Skip to main content

Full text of "Watch and clock making"

See other formats


This is a digital copy of a book that was preserved for generations on library shelves before it was carefully scanned by Google as part of a project 

to make the world's books discoverable online. 

It has survived long enough for the copyright to expire and the book to enter the public domain. A public domain book is one that was never subject 

to copyright or whose legal copyright term has expired. Whether a book is in the public domain may vary country to country. Public domain books 

are our gateways to the past, representing a wealth of history, culture and knowledge that's often difficult to discover. 

Marks, notations and other maiginalia present in the original volume will appear in this file - a reminder of this book's long journey from the 

publisher to a library and finally to you. 

Usage guidelines 

Google is proud to partner with libraries to digitize public domain materials and make them widely accessible. Public domain books belong to the 
public and we are merely their custodians. Nevertheless, this work is expensive, so in order to keep providing tliis resource, we liave taken steps to 
prevent abuse by commercial parties, including placing technical restrictions on automated querying. 
We also ask that you: 

+ Make non-commercial use of the files We designed Google Book Search for use by individuals, and we request that you use these files for 
personal, non-commercial purposes. 

+ Refrain fivm automated querying Do not send automated queries of any sort to Google's system: If you are conducting research on machine 
translation, optical character recognition or other areas where access to a large amount of text is helpful, please contact us. We encourage the 
use of public domain materials for these purposes and may be able to help. 

+ Maintain attributionTht GoogXt "watermark" you see on each file is essential for in forming people about this project and helping them find 
additional materials through Google Book Search. Please do not remove it. 

+ Keep it legal Whatever your use, remember that you are responsible for ensuring that what you are doing is legal. Do not assume that just 
because we believe a book is in the public domain for users in the United States, that the work is also in the public domain for users in other 
countries. Whether a book is still in copyright varies from country to country, and we can't offer guidance on whether any specific use of 
any specific book is allowed. Please do not assume that a book's appearance in Google Book Search means it can be used in any manner 
anywhere in the world. Copyright infringement liabili^ can be quite severe. 

About Google Book Search 

Google's mission is to organize the world's information and to make it universally accessible and useful. Google Book Search helps readers 
discover the world's books while helping authors and publishers reach new audiences. You can search through the full text of this book on the web 

at |http: //books .google .com/I 

Manuals of Technology. 


Prof, Ayrton^ F,R.S,^ and R. Wormell^ D,Sc,y M.A, 

Watch and Clock Making. 






[all bxohts bbss&vbd.] 




P E E P A C E. 


As so many books tad lately been published on Watch 
and Clock Making, it was with some hesitation that I 
accepted the task of writing this manual However, 
on further reviewing the horological literature in our 
language, it appeared so scant in comparison with the 
importance and interest of the subject, and so firm 
"^as my conviction that the entire absence of any 
system of technical education was the chief cause 
of the decadence of the trade in this country (which 
has been so much felt during the last few years), that I 
was induced to undertake the work from the desire of 
promoting the educational movement now going on, 
and of assisting the rising generation of watchmakers 
to overcome those difElculties which had proved such 
stumbling-blocks to myself and others at the outset 
of our careers. 

This volume is intended not only as a text-book for 
technical classes, but is designed also as a book of 
reference for the practical workman. 

In the introductory sections I have summarily traced 
the progress of the watch and clock trade down to 
the present time, and have endeavoured to show by 


what very gradual steps real improvements take place 
or are appreciated. In the remainder of the book I have 
combined the theoretical treatment with the practical 
in such a manner as will, I trust, prove both interest- 
ing and instructive to the student, although from the 
limited space at my disposal the whole can be regarded 
only as a compendium of an almost inexhaustible sub- 
ject. The principal drawings have been done by my 
son, who has also generally assisted me in the prepara- 
tion of the manual 

Want of space has likewise prevented me from 
adding a few chapters on Watch-Jobbing, as was my 
original intention. A careful study of the practical 
portions of the book will, however, enable any work- 
man to repair or replace injured or broken parts of the 
mechanism in the proper way, and the explanation of 
the theory and, action of the various escapements will 
direct him in detecting any error or wrong propor- 
tions in any given escapement—the most important and 
difficult task the watch-jobber has to perform. 


20, Myddeltm Square, E.C, 




L Time. 2L A Sidereal Day. 8. A Solar or Ciyil Day. 4. Hean Solar 
Time. 5. Tracing a meridian. 6. Snn-diala and ClepsvdriB. 7. 
Sand^GIaases, Candle Clocks, and Orreries. 8. Wheel Clocks. 9. 
De Wyck's Clock. 10. Isochronism of the Pendnlnm. 11 The 
English Clock Trade. 12. The Westminster Clock .... 1 



13. The Earliest Watches. 14. Hooke's Law. 15. The Clockmakers* 
Company. 16. Bepeating Work. 17. Jewelling and the Horizontal 
Escapement. 18. Harrison's Marine Timekeei>er. 19. Ifudge's 
Chronometer. 20. Arnold's Escapement. 21. The Cylindrical 
Spring. 22. EamshaVs Balance. 28. Attziliary Compensation. 
24. Improved Chronometer Balances. 25. Hartnnp's Balance. 28. 
Heroer s Balance. 27. Compensated Watches. 28. Ships' Chrono- 
meters. 29. Exiglish Watches. 80. Apprentices and Pupils. 81. 
Machine-made ItoTwnents 17 



82. The Movement Trade. 83. The Full Plate. 84. Mr. Wvcherle/s 
Improvements. 85. Movement Plates. 86. Screw Making. 87. 
Pimoxi Making. 88. Wheel Teeth. 89. Pillar Plates ... 87 



40. Drivers and Followers. 4L Helical Teeth. 42. Construction of 
Wheel Teeth 41 



43. Trains. 44. Motion Wheels • • 53 



45. Tlie Fnsee. 46. Fusee Stop. 47. Mainsprings. 48. Numher of 
Turns in the Spring. 49. Spring Attachments. 50. Going-harrel 
and Stop Work. 51. Mainspring Making 63 




52. Chronometer rinishing. 53. Hooking in the Spring. 54. Planting 
the Wheels. 55. Stop and Spring. 56. Adjosting-rod. 57. Finish- 
ing the Wheels and Pinions. 58. Jewelling. 59. Spotting the 
Plates. 60. Watch Finishing. 61. Testing the Wheels. 62. Eun- 
ning in the Wheels. 63. The Top Plate. 64. The Barrel Arbor. 
65. HooMng in the Spring. 66. Adjusting the Chain. 67. Pivoting. 
68. Motion Work. 69. Screws. 70. Suggestions as to Finishing. 
71. Applying the Balance Spring 80 



72. Watch Examining. 73. Fitting the Hands. 74. Attaching the Dial. 
75. General Revision. 76. Fixing Movement in Case . .103 



77. Objections to Jewelling. 78. Stones for Jewelling. 79. Hole 
Making. 80. Jewelling. 81. Snailing. 82. GUding . . . .110 



88. What is an Escapement. 84. Yerge Escapement. 85. Hori- 
zontal Escapement. 86. To Plan a Horizontal Escapement. 87. 
Parts of the Horizontal Escapement. 88. The Cylinder. 89. 
Making a Cylinder 123 



90. Principle of the Escapement. 91. Angular Measurement of Lift. 
^ 92. Running of the Wheel. 93. Duplex Escapement. 94. Escape 
Wheel. 95. The Escaping Angle. 96. Wheel Teeth. 97. Pallet 
Roller 137 



96. Early Forms of the Escapement. 99. Present Form. 100. Lift. 
101. Long and Short Detents. 102. System of Manufacture. 103. 
Marine Chronometer Escapement. 104. Pivots. 105. Locking 
Stone. 106. The Detent 155 



107. Preparing the Different Parts. 108. Spring Detent. 109. Banking 
Piece. 110. The Unlocking Spring 169 





111. Double Boiler Escapement. 112. The Back Lever. 113. The 
Detached Lever. 114. The Crank Lever. 115. Table Boiler. 116. 
Two-pin Escapement. 117. The Doable Boiler. 118. Cole's Be- 
silient Escapement. 119. Cole's Bepellent Escapement. 120. The 
Club-tooth Lever. 121. Straiffht-liue Escapement. 122. Fallets 
with Circular nnd Equidistant Lockings. 323. Caliber. 124. Steady, 
pins. 125. Sinking the Escape-wheel. 126. Balance. 127. The Im> 
pulse Pin. 128. The Guard Pin. 129. The Pallet Staff. 130. Lever 
and Pallets. 131. Escape-wheel ........ 176 



182. Invention. 133. Hooke's Law. 134. The First Spiral Spring!. 
185. Different Forms of Spring. 136. Isochronism of Balance 
Spring. 137. Isochronism and Length of Spring. 138. Comi>ensa- 
tion. 139. Flat Springs. 140. Correcting the Spring. 141. Glass 
Springs. 142. Palladium Springs. 143. Demagnetising Steel Work 205 



144. Testing Spring Wire. 145. To make the Winder. 146. Polishing 
the Spring. 147. Breguet Spring. 148. Apnlying the Spring. 149. 
Pinning on the Spring. 150. Caution about Sreguet Spring. 151. 
Chronometer Spring. 152. Blueing Springs. 153. Applying Spring 
to Chronometer 217 



154. "Middle Temperature Error." 155. Berthoud's Experiments. 
156. Effect of Temperature upon SpringiS. 157. Safe Bauge of 
Temperature. 158. Daily Bate of Chronometer. 159. Duration of 
Test. 160. Airy's Experiments. 161. Hartnup's Suggestions for 
Checking the Error. 162. Adjustment of Chronometers. 163. The 
Error of the Compensation Balance 227 



16k Hardy's Balance. 165. Eiffels or Molineux's Balance. 166. Dent's 
Balance. 167. Hartnup's Balance. 168. Poole's Balance. 169. 
Loseby's Balance. 170. EuUberg's Balance. 171. EuUberg's Im- 
provea Balance. 172. Meroer's Balance. 173. Airy's Compensation 
Bar. 174. Chronometer Balances. 175. Chronometer Trials . . 242 



176. Caron's Keyless Watch. 177. Other Contrivances. 178. Prest's 
- Keyless Work. 179. Swiss Keyless Work. 180. Booking Bar Key- 
less Work. 181. Fusee Keyless Work. 182. KuUberg's Fusee 
. Keyless Work. 183. Chalfont's Fusee Keyless Work. 184. The 
various forms of Keyless Work ........ 258 




185. The First Pendnlom Clock. 186. Isochronifinn of the Pendulum. 
187. The Circular Error. 188. Time of One Swing of Simple Pendu- 
lum. 189. The Centre of Oscillation. 190. Su8i>en8ion of Pendu- 
lum. 191. Pendulum Cock. 192. Pendulum Bobs. 193. Barometric 
Error. 194. Begnlation 274 



195. Bat'O of Ezmnsion of Metals. 196. Herourial Compensation 
Pendulum. 197. The Gridiron Pendulum. 198. Wood and Zinc, 
or Wood and Lead, Pendulum. 199. Zinc and Steel Compensation 285 



200. Becoil Escapement. 201. Pin Wheel Escapement. 202. Dead- 
beat (Graham) Escapement. 203. Conditions of Isochronous Vibra- 
tion. 204. Erodsham's Tables 292 



206. Double Three-legged Gravity Escapement. 206. Four-l^ged 
Gravity Escai>ement. 207. Electrical Clocks. 208. Jones's SyBtem 
of Controlling Clocks. 209. Eitchie's Electrical Clock Escapement . 308 



210. Various kinds of Clocks. 211. Bight-day Clocks. 212. The Strik- 
ing Train. 218. Quarter Clocks. 214. Astronomical Clocks or 
Begulators. 215. Turret Clocks 319 



fig. 1. — Traiuit Instrument 3 

Fig. 2. — Generating Oircles for Epicycloid and Hypocycloid . 46 

^. 3.— Bevelled Wheels (diagram) 62 

Fig. 4. — ^Fnsee and Barrel with Chain attached .... 63 

Fig. 5.— The Fusee 64 

Fig. 6. — A, Ordinary Going Barrel and Stop Work ; B, Improved 

€k>ing Barrel ; o, Stop Piece ..... 76 

Fig. 7. — Adjusting-rod , . 83 

Fig. 8.— Old-fashioned Swiss Stud 102 

Rg. 9.— Jewel Hole 116 

Rg. 10. — Verge Escapement 125 

Fig. IL — Horizontal or Cylinder Escapement .... 128 

Hg. 12. — Horizontal Escapement (diagram) .... 130 

Fig. 13. — ^Tool for Running in Cylinder 134 

Hg. 14.— Tool for measuring Heights 136 

Fig. 15. — ^Duplex Escapement : — I. Plan and elevation ; II. 

Action of impulse 138 

Fig. 16. — Angular Measurement of lift 140 

Fig. 17. — ^Diagram — Proportions of Wheel and Pallet . . 142 

Fig. 18.— Ditto ditto ditto ... 143 
Fig. 19.— Ditto ditto ditto . . .144 
Fig. 20. — ^Action of Escape Wheel, with Boiler, &c., of correct 

proportions 147 

Fig. 21. — ^Action of Wheel with too laige a Boiler . • •148 

Fig. 22. — ^Action of Wheel with too small a Boiler . . • 149 

Fig. 23. — Chronometer Escapement (plan) . . . • • 156 

Fig. 24. — Chronometer Escapement (elevation) .... 157 

Fig. 25. — Diagram — Second Order of Lever .... 159 

Fig. 26.— Tool for measuring Heights ...... 164 

Big. 27.— Conical Pivot 165 

Fig. 28.— Pivot with Enlarged End 166 

Fig. 29.— Double Roller Lever Escapement :— I. Plan ; IL Eleva- 
tion 177 

Fig. 30.— Rack Lever 179 



Fig. 31.— Crank Level 181 

Fig. 32. — Two-pin Lever Escapement 183 

Fig. 33. — Cole's Besilient Escapement 186 

Fig. 34. — Straight-line Lever Escapement, with Club Teeth . 189 

Fig. 35.— Circular Pallets 191 

Fig. 36. — ^Pallets with Equidistant Lockings .... 192 

Fig. 37. — Equidistant Locking v. Circular Pallets (diagram) . 193 

Fig. 38. — ^Pallets, with Semi-equidistant Lockings . . . 194 

Fig. 39. — Guard Pin in Lever (diagram) 201 

Fig. 40. — 1, Ordinary Volute ; 2, Cylindrical Helical ; 3, Breguet 

Spring 208 

Fig. 41, — 1, Box-cover ; 2, Spring-box ; 3, Screw for Fixing 

Wire ; 4, Plan of Winder-pivot ; 5, Winder . . 218 

Fig. 42.— To polish Outer Sides of Coils 219 

Fig. 43. — ^To polish Inner Sides of Coils 220 

Fig. 44.— Winder for Breguet Spring 220 

Fig. 45.— Diagram 230 

Fig. 46. — Ordinary Compensation Balance with sliding weights . 243 

Fig. 47.— Hardy's Balance 244 

Fig. 48.— Eiffe's or Molineux's Balance 245 

Fig. 48a.— Eiffe's or Molineux's Balance 245 

Fig. 49.— Dent's ^Balance 247 

Fig. 50. — Hartnup's Balance 248 

Fig. 51. — Poole's Auxiliary 250 

Fig. 62.— Loseby's Balance 252 

Fig. 53.— Kullberg's Flat-rim Balance 253 

Fig. 54. — Eidlberg's Improved Balance 254 

Fig. 55. — Mercer's Balance 256 

Fig. 56.— Brest's Keyless Work 259 

Fig. 57.— Swiss Keyless Work 261 

Fig. 58.— Rocking Bar Keyless Work 264 

Fig. 59.— Fusee Keyless Work 268 

Fig. 60.— Kullberg's Fusee Keyless Work 271 

Fig. 61.— Chalfont's Fusee Keyless Work 273 

Fig. 62. — Mercurial Compensation Pendulum .... 286 

Fig. 63.— Recoil Escapement 294 

Fig. 64.— Pin Wheel Escapement 296 

Fig. 65. — Dead Beat (Graham) Escapement .... 298 

Fig. 66.— Double Three-legged Gravity Escapement , , . 309 

Fig. 67. — Jones's System of Controlling Clocks .... 314 

Fig. 68. — Ritchie's Electrical Clock Escapement .... 317 

Watch and Clock MAKma 



1. Time. — Two fundamental ideas form the basis of 
all investigations in physical phenomena : these are 
duration and extension, or time and space. Both are 
conceived by us as unbounded, continuous, and indefi- 
nitely divisible. We can conceive of them existing 
"without matter, for if the whole material universe were 
destroyed, time and space would remain. Neither, how- 
ever, could be measured without the existence of matter. 
Portions of space can be considered only by reference to 
material things, and portions of time only by means of 
motion. Measured time may be defined as the per- 
ceived number of successive movements. In other words, 
time is duration, as set out by certain periods and 
marked by certain measures and epoch& 

The measures of time are days, months, years, and 
cycles ; and, for the convenience of the ordinary routine 
and business of life, man has from time immemorial sub- 
divided the natural periods into shorter, and what may 
be called mathematical, periods, such as hours, minutes, 
and seconds ; for the recording and measuring of which, 
the different machines used in horology have been in- 
vented and constructed. 

Amongst the primitive nations the sub-divisions of 
the day were always more or less marked. The ancient 
.Komans divided it into four parts, while in most 
civilised countries it is now divided into twenty-four 



hours, which are each sub-divided into sixty minutes, 
each of which contains sixty seconds. Our clocks and 
watches are n)arked for twelve hours only, which are 
twice counted. 

The Italians reckon from one to twenty-four (which is 
also the mode of reckoning in our observatories with as- 
tronomical clocks), while the Chinese divide the day into 
twelve hours only, each of which is equal to two of ours. 

A day is the time of the complete rotation of the earth 
on its axis, either with reference to the stars or to the 

2. A Sidereal Day. — The interval between successive 
transits of the same star is termed a sidereal day. It is 
the true period of the earth's rotation. It is ascertained 
with great exactness by the ordinary observations of astro- 
nomers, and is, as far as we know, invariable. It is the 
standard by which all clock time is finally regulated, and 
is daily taken at observatories by observing particular 
stars by means of a transit instrument : i,e., a telescope 
fixed at right angles across a stiff" axis with pivots at the 
ends, which is laid horizontally east and west on firm 
supports, thus allowing the telescope to move only in the 
plane of the meridian (Fig. 1). It has wires stretched 
vertically across the focus, and it is by the sun or par- 
ticular stars being observed to pass this meridian that 
we measure whole days. 

(For a full description of the transit instrument and 
its uses, see " A Treatise on the Transit Instrument as 
applied to the Determination of Time," by Latimer 
Clark, M.I.C.R, &c., and "Transit Tables for 1884, 
giving the Transit of the Sun and of certain Stars for 
every Day in the Year, computed from the * Nautical 
Almanac ' for Popular Use,'' by the same author. Lon- 
don : Alfred J. Frost, 6, Westminster Chambers, Victoria 
Street, S.W.) 

Mr. Clark, in his treatise, gives the following instruc- 
tions for using the transit instrument : — 

" The direct result of a transit observation anywhere 

made is tlie local time at that spot, but by adding a cer- 
tain conBtant, whicli never varies, we convert this into 
Greenwich time. This constant depends on our longitude 

east or west of Greenwich. The fir^t requisite is, 
therefore, to find to what exteot our station is either 
east or west of Oreenwicli, or, ia other words, to take our 
longitude by means of a map. Let us, for the present, 
assume our station to be at Marlow, which Is 3' 6" in 
time west of Greenwicli. 


"The Transit Tables give lis, with minute accu- 
racy, the time at which the sun passes across the 
meridian of Greenwich, i.e.j the time at which the sun 
is due south at Greenwich, for every day in the year. 
Now, since our longitude is 3' h" west of Greenwich, the 
Bun will pass every day over our meridian at Marlow 
3' 5" after it has passed over the meridian at Greeewich. 
We have thus merely to add 3' 5" to the times given 
in the table, and we have for every day in the year the 
instant at which the sun will be due south at Marlow, 
and this is the time which should be shown by our clock 
at that instant, if it is right. 

" The sun's motion in the heavens is so slow that it is 
scarcely perceptible to the unassisted eye, but when seen 
through a telescope, the motion, as well as its size, are so 
much magnified, that the instant of its passage over tlio 
wires of the telescope can be observed within the frac- 
tion of a second of time. As an example of its use, we 
observe by our watch on March 1 that the sun passed 
over the middle wire of our telescope at 14' 32" past 
12. We find by our tables that on this day the 

Sun souths at Greenwich at . 
Add for our longitude 







But the time shown by our watch "was 




Difference • 




"It is thus evident that our watch was 1' 3.82" too 

" The tables give in a similar manner for every evening 
the instant of southing of most of the principal stars. 
Their rapid motion across the wires of the telescope in- 
sures great accuracy in determining the time of their 
transit, and as the tables give not only the time of their 
southing, but also the altitude of the stars at the time 
the^ pass the meridian, their identification is insured by 


means of the graduated circle attached to the instrument 
on which altitudes are measured. As an example of the 
process, we observe that on March 2 the bright star 
a Tauri crossed the centre wire of our telescope at 
5® 50' 50" according to our watcL We find in our tables 
that on this day a Tauri southed at 

H. M. a 

Greenwich at 6 47 31.7 

It will therefore south at Marlow on 
account of longitude later . . 3 5 

6 60 36.7 

By our watch the transit occurred at . 6 50 60 

Difference . . 13.3 

"Our watch was evidently therefore 13' 3" too fast." 

The duration of a sidereal day is 23 hours, 56 
minutes, 3 seconds. 

3. A Solar or Civil Day is the time between the tran- 
sits of the sun over the meridian on two successive days ; 
but as the sun revolves relatively in the same direction 
as the eai-th rotates [of course, strictly speaking, it is the 
earth which revolves round the sun; it is only apparently 
that the sun moves round us], the sun requires nearly 
3 minutes, bQ seconds on the average longer than any 
particular star to bring him up to the same meridian on 
every successive day ; there is therefore one more actual 
or sidereal day in the year than there are solar days. 

There are two causes which render the lengths of 
solar days unequal The first is the variable velocity of 
the eaiiih in its orbit, which is an ellipse, having the 
sun not in the centre but at a focus, and in which the 
earth moves slowest when it is farthest from the sun, 
or in apogee, and quickest when nearest the sun, or in 
perigee. And the second is the fact that the earth does 
not move (or, which is the same thing, the sun does not 
seem to move) in the plane of the earth's equator, but in 
a plane inclined to it. The apparent path of the sun in 
the heavens is called the ecliptic^ and the second of the 


causes making the solar day of variable length may 
therefore be described as the obliquity of the ecliptic. 

4. Mean Solar Time. — A mean solar day is the 
average length of all the solar days in the year, which 
is 24 hours. Conceive an imaginary sun moving uni- 
formly in a circle in the plane of the earth's equator so 
as to coincide with the real sun when their paths (the 
ecliptic and celestial equator) cross each other, and there- 
fore so that the year as marked by the imaginary and 
real suns is exactly of the same length. Then the 
imaginary sun will mark the mean solar day. The dif- 
ference between true and mean time is called the equa- 
tion of time. 

There are only four days in the year when the 
apparent and mean time are the same, and the equation 
of time nothing. These are December 24th, April 16th, 
June 15th, and August 31st. Between December 24th 
and April 15th, and between June 15th and August 31st 
the apparent is always before the mean time, whilst in 
the remaining intervals it is later; the almanacs generally 
give it " clock before *' and " clock after " sun. 

5. Tracing a Heridian. — The equator, and any circle 
on the earth parallel to the equator, contains 360 degrees, 
each of which is divided into 60 minutes, which are 
again divided into 60 seconds. Circles passing through 
poles dividing the equator and all parallel circles into 
degrees are called lines of longitude. Places in the same 
line of longitude have their noon or midday at the same 
instant; hence these lines are called meridians of 

As the earth rotates once in 24 hours and there are 
360 degrees of longitude, each degree of longitude = 
4 minutes ; so that for ascertaining the local time at any 
place it is only necessary to add or subtract for the 
distance east or west of the meridian that is chosen for 
referenca The meridian of Greenwich is the one now 
almost universally used as the initial meridian. It is 
important to know how to find the directions of the 


lueridian at any spot Ferguson gives the following 
directions for tracing a meridian : — 

" Make four or five concentric circles, a quarter of an 
inch from one another, on a flat board about a foot in 
breadth, and let the outmost circle be but little less than 
the board will contain. Fix a pin perpendicularly in the 
centre, and of such a length that its whole shadow may 
fall within the innermost circle for at least four hours in 
the middle of the day. The pin ought to be about the 
eighth part of an inch thick, with a round blunt point. 
The board beinff set exactly level in a place where the 
sun Bhines, sup^ from eight in the morning till four in 
the afternoon, about which houi-s the end of the shadow 
should fall within all the circles ; watch the times in the 
afternoon, when the extremity of the shortening shadow 
just touches the several circles, and there make marks. 
Then in the afternoon of the same day, watch the 
lengthening shadow, and when its end touches the several 
circles in going over them make marks also. Lastly, 
with a pair of compasses, find exactly the middle points 
between the two marks on any circles, and draw a 
straight line from the centre to that point, which line 
will be covered at noon by the shadow of a small upright 
wire, which should be put in place of the pin. The 
reason for drawing several circles is, that in case one 
part of the day should prove clear, and the other part 
somewhat cloudy, if you miss the time when the point of 
the shadow should touch one circle, you may perhaps 
catch it in touching another. The best time for drawing 
a meridian line in this manner is about the middle of 
summer, because the sun changes his declination slowest 
and his altitude fastest in the longest days.'' 

As a substitute for the transit instrument, Bloxam 
invented and patented an instrument called a dipleido- 
scope, which consists simply of three plates of glass, 
put together at their edges and forming a hollow 
triangular prism ; two of these plates reflect inwards, 
while the front is plain, and, if placed at noon so that 


there is but one image of the sun reflected by the 
incident rays from these reflectors, it makes a very 
good meridian instrument, with the advantage of en- 
abling an observation to be taken when the ordinary 
shadow instrument would not be sufficiently distinct.. 

In taking observations with these instruments with a 
view to correct clock time, it is necessary to take into 
account the difference between mean solar or true time 
and eqtial or clock time; the simplest means of doing 
which is to consult a good almanac that gives the sidereal 
time for each noon in the year. 

A further correction for longitude would have to be 
made if Greenwich mean time were required by those 
who are not in that meridian, by adding or subtracting 
as described on pages 4 and 5. 

6. Sundials and ClepsydrsB. — The first gnomon or 
sundial of which we have any historical notice is that 
of King Ahaz, about 740 B.C. It is mentioned in 
2 Kings XX. 11 : "And Isaiah the prophet cried unto 
the Lord : and he brought the shadow ten degrees back- 
ward, by which it had gone down in the dial of 

The ancients used hemispherical dial-plates, with the 
radius which throws the shadow running north and 
south ; those now in use are flat, with the edge of the 
shadow radius forming an angle with the horizon equal 
to the latitude in which they are situated or parallel to 
the earth's axia Although it was possible to tell the 
time by a mathematically adjusted sundial to within a 
few minutes, they were in a great measure superseded at 
a very early date by clepsydrae and sand glasses, though 
in the face of our want of knowledge of the exact dates 
of their invention, it is not easy to say which of them 
was in existence first ; it is cartain, however, that both 
were known and in use at a very remote period. They 
were found even in Britain by Julius Csesar, who observed 
by them the difference in the lengths of the days and 
nights of tliis country and of those of Italy. They were 


introduced, along with other arts and sciences, to Europe 
from the East, and it is probable that the intercourse 
between the people of this country and the Phoenicians 
may account for our thus early possessing them. 

Toothed wheels are said to have been first applied to 
clejisydrse by Ctesibius, a native of Alexandria, about 
250 B.C. 

Amongst the various descriptions given by ancient 
writers of the different forms of clepsydrae, I see none 
of any method for regulating or counteracting the 
irregular pressure of the water consequent on its running 
away or through the varying density of the atmosphere. 
The most usual form seems to have been a graduated 
vessel containing a float, into which water dropped from 
another vessel, the float as it rose indicating the time of 

7. Sand Glasses, Candle Clocks, and Orreries. — The 
sand clocks or glasses seem to have been very similar 
to an ordinary hour-glass of the present day. 

Candle clocks were a later method of marking time. 
King Alfred the Great is said to have used them, but 
whether he was the inventor or not is doubtful. 

It is rather remarkable that long before the invention 
of wheel clocks, planetariums or orreries were well 
known. The first modem planetarium in England was 
one made for Lord Oirery, whose name has been since 
generally given to these machines, and that our fore- 
fathers looked upon the planetary motions as the true 
measure of time may be proved by the fact that the first 
clocks constructed nearly always showed various astrono- 
mical phenomena, in addition to showing or striking the 
hours of the day. 

8. Wheel Clocks. — ^Yery little is known about the 
first invention of wheel clocks, no two writers seemingly 
agreeing as to the exact period of their introduction. 
Although some historians assign to it such an early date 
as even the sixth century, it is not very probable that 
they are correct, the words horologium, horologe, &c., 


having been applied indiscriminately in old writings to 
any machines for measuring thne. 

It is pretty certain, however, that clocks driven by 
weights and striking automatically existed in the eleventh 

Gerbert, afterwards Pope Sylvester II., is said to 
have made a clock at Magdeburg in the year 996 ; but 
there seems to be a considerable difference of opinion as 
to whether it was a clock at all, and Dillmar of Merse- 
burg states that it waa only a kind of sundial. 

The oldest clocks in England were that of St. PauFs 
Cathedral, London, and one at Westminster, which latter 
was paid for out of a fine imposed by Edward I., in 
the year 1288, upon Sir Kalph de Hengham, Chief 
Justice of the King's Bench, for corrupt practices. It is 
said to have remained until the time of Queen Elizabeth. 
The former clock is mentioned in the " Compotus 
Bracerii" of St. Paul's for the year 1286, where the 
allowances to the clockmaker, " Bartholomo Orologiario," 
are mentioned. 

9. De Wyck's Clock. — From these dates the manu- 
facture of clocks would appear to have become a settled 
industry in this country, although the fii-st authen- 
tic description we have of the interior of any wheel 
clock is that of one made by Henry de Wyck, a 
German, for Charles Y. of France in 1379, which has 
been not inaptly styled the "parent. of modem time- 
keepers," since, except that it had no pendulum and only 
an hour hand, it was very like, in principle, the clocks 
of the present day. It consisted of a train of wheels 
driven by a weight, and had a vertical or verge escape- 
ment with a vibrating balance, but no spring ; the 
balance instead of being shaped like a fly-wheel was in 
the form of a T, upon the two thin projecting arms of 
which concentric notches were cut. Two small regulat- 
tag weights were suspended from the arms, and it was by 
shifting these from notch to notch, to or from the centre, 
that the clock was made to go faster or slower as required. 


A description of De Wyck's clock by Berthoud in his 
" Histoire de la Mesure du Temps par les Horologes," 
1802, states the number of teeth in the escape wheel 
or wheel of rencounter to have been 30, which must 
be a mistake, as it is impossible for the pallets of a crown 
wheel or verge escapement to " escape " with an even 
number of teeth in the wheel. 

As this was the general mode of construction adopted 
in the fourteenth century, the pendulums found in the 
old church clocks of Exeter, Peterborough, and else- 
where, must have been added long after their original 
construction. This was the case also with the minute 
hands later on, when the '* works" were utilised; and 
the dials for showing the time were in many instances 
added to the original clocks, where these had at first 
been made to strike only. 

Some of the old clocks "vith only an hour hand are 
still in existence in this country, as that of St. Margaret's, 
Westminster; and one of the dials of St. Paul's clock 
was described in 1844 as being the largest in this country 
furnished with a minute baud. 

As the train of wheels in a watch or clock is driven 
by a weight falling, or by the uncoiling of a spring, it is 
necessary that the power should be allowed to expend 
itself only in equal quantities, otherwise the motion 
given to the train would be a gradually accelerating one 
in accordance with the law of gravity. 

It is obvious, therefore, that unless some measure 
were adopted for controlling and regulating the power, 
the mechanism would speedily run down and exhaust it. 

The usual method of regulating the power of the prime 
mover has always been by controlling the motion of the 
fastest wheel of the train : this is called the wheel of 
rencounter or " escape " wheel, because it is allowed to 
escape from the controlling stops, or pallets, as they are 
called, only one tooth at a time. 

The earliest escapement of which we know anything 
is that of De Wyck's clock. This was the same in principle 


as the vertical escapement of the present day ; that is, 
it had two pallets on a staff or axis at right angles to 
one another, and which by receiving the impulse alter- 
nately from either side of the crown wheel into which 
they worked, allowed one tooth to "escape" at every vibra- 
tion of the balance which was attached to their axis ; as 
one tooth "escaped," the opposite pallet received the im- 
pulse from the next tooth on the other side of the wheel. 
The balance had, however, no spiral or balance spring, 
and so its vibrations must have varied in length (and in 
the length of time occupied by them, as they could 
not have been isochronous), according to the strength or 
weakness of the impulse. 

A clock having a similar escapement to the above 
was exhibited in 1877 at the Special Loan Collection 
of Scientific Apparatus, South Kensington. It was of 
Swiss manufacture, and supposed to have been made in 
the year 1348. 

10. Isochronisin of the Pendulum. — The discovery 
of the isochronism of the pendulum is attributed to 
Galileo, who observed that a chandelier swinging in 
a church at Florence performed the vibrations of 
the long and short arcs in the same time. It was 
he who first conceived the idea of applying the 
pendulum to a clock. This clock his son Vincenzo 
made in 1649, but, according to the account given of 
it by Galileo's disciple Viviani to Prince Leopoldo de' 
Medici, it was not by any means perfect, and it was 
while engaged in perfecting it that Vincenzo died, on the 
16th of May, 1649. 

Huygens, the Dutch philosopher, was the first to in- 
vestigate thoroughly the mathematical theory of the 
pendulum. He also, along with Dr. Hooke and others, 
claimed the credit of the original discovery, although, 
apart from its application and from Huygens' investiga- 
tions, there does not appear to be much merit in its 
mere discovery. Sir Edmund Beckett, referring to the 
discovery of Galileo, says, " When we consider the vast 


number of pendulums of various kinds swinging about 
the world, it certainly is difficult to imagine that no- 
body ever made that observation before the sixteenth 

Huygens discovered and established the fact that the 
oscillations of a pendulum, in order to be isochronous, 
should describe a cycloidal curve — i,e., a curve traced 
by a point in the circumference of a circle rolling upon a 
straight line. And in order to compel the pendulum bob 
to travel in this curve, he invented the cycloidal cheeks 
between which the spring was suspended. Although 
this was thought to be a very great improvement at the 
time, the idea was soon given up, it being found that 
clocks constructed without the cheeks went better than 
those constructed with them. The theory would be true 
enough of what is called a simple pendulum — i.e., a 
mathematically perfect pendulum in which there is no 
weight but at the centre of oscillation — ^but as in prac- 
tice there is no such thing, it is found that any advantages 
which a pendulum so suspended may theoretically possess 
over an ordinary one, are more than counterbalanced 
by drawbacks in the shape of friction against the cheeks, 

But, as to perform isochronous vibrations a pendulum 
ought not to deviate from its true theoretical path (which 
is a cycloidal arc), so any variation of the arc actually 
described by an ordinary pendulum (which is a segment 
of a circle) alters its time of oscillation. 

This error, which is called the circular error (being 
the diflerence between the time occupied by a pendulum 
oscillating in a cycloidal and in a circular arc), is best 
overcome, or rather prevented to a great extent, by using 
a long pendulum describing small equal arcs; and, in 
order that the arcs described should be kept equal, 
the impulse must be equal too. This equality of im- 
pulse has been best secured by the adoption in modem 
clocks of what are called " dead beat " escapements, and, 
in the case of exposed clocks, of the " gravity " escape- 


ment, a description of which will be found in its proper 

As the invariable length of the pendulum is such an 
important factor in the performance of the clock, it was 
natural that its compensation for varying temperatures 
should have been one of the first great problems early 
horologists had to solve ; in fact it is a subject which has 
occupied the minds of clockmakers even to the present 

The principle of Compensation adopted has nearly 

always been the construction of the pendulum with two 
or more metals of different expansibility, so arranged 
that the position of the centre of oscillation shall remain 
approximately unaltered. 

The most successful inventions have been, for regula- 
tors and house clocks, tbe mercurial and gridiron pen- 
dulums, and for large turret clocks, the zinc and iron 
compensation which, while being as effective as the mer- 
curial, is a good deal cheaper. 

11. The English Clock Trade, with the exception of 
the trade in turret clocks, has never reached to any degree 
of importance, and could never at any time have been 
classed with our large national industries ; while of late 
years the trade in house clocks may be said to have 
almost entirely left us. The French have taken what 
may be called the drawing-room clock trade out of our 
hands, the proverbially elegant taste of that nation, 
aided by a better artistic education and a more ener- 
getic and intelligent application of it amongst those 
interested in the manufacture, having enabled makers to 
produce clocks which in quality and price have entirely 
superseded the same class of English goods. English 
makers, having unfortunately used no effoi-ts to supply 
obvious public requirements either in the way of 
cheapness or elegance, suddenly awoke to the fact 
that their trade was gone ; so that with the excep- 
tion of turret clocks, and of 6rst-rate regulators and 
quarter clocks, there are now only a few eight-day 


kitchen and school clocks made in England, and even 
the first of these industries is hard pressed by the 
cheaper importations from Holland, Belgium, Germany, 
and America. The quarter clock, striking, and gen». 
rally repeating the hours and quarters, is a distinctly 
English production, and has never been rivalled by any 
foreign-made clock in its own peculiar province, that 
of a hall clock. Even in these days of low prices and 
keen competition, it has generally maintained its posi- 
tion, quality, and even price, which it commands now 
as much as it did fifty years ago. It is made mostly 
in London by the best clbckmakers, and its handsome 
case and sonorous gongs render it a pleasing characteris- 
tic of an English hall. It is made sometimes to chime 
on gongs and sometimes on bells, and often on both, when 
they can be alternated at pleasure by setting a hand 
which is on the dial for this purpose; it can also be 
rendered silent by moving a pin fixed at the side of 
the dial. Of late years a large number of imitation 
quarter and other old English clocks have been imported 
from France, and sold as English-made clocks, and the 
public has doubtless largely suffered by the deception. 
The imitation article, however, can eaaUy be detected in 
the case of large clocks by noticing the distance from 
one another of the winding squares, or of the holes in 
the diaL In an English clock, in which the size of the 
movement corresponds to that of the case, these winding 
squares are some distance apart, but in the foreign imi- 
tation they are almost always close together, the size 
of the movement bearing but small proportion to the 
large and showy exterior. The clocks imitated are 
usually the quarter clocks and the early English clocks, 
with porcelain dials and antique figuring. 

Regulators or astronomical clocks are still made here, 
and also exported on a small scale, and, save some 
cheaper attempts at the same thing imported from 
Austria and Germany, have no foreign rival. And it 
is probable they will hold their own with the quarter 


clocks, since there is not suflficient demand for them to 
make their wholesale manufacture pay. 

The turret clock trade is carried on necessarily on 
the factory system, and, like most large English me- 
chanical work, is not likely soon to be beaten by that of 
any other country, 

12. The Westminster Clock. — The construction of 
the great Westminster clock by the eminent scientist 
and horologist. Sir Edmund Beckett, caused a revolution 
in this branch of the clock trade. The design upon 
which it was made, which was somewhat unfavourably 
criticised at first and compared by some of the old school 
of clockmakers to a mangle, is now accepted as a model 
after which all the best turret clocks are fashioned with 
such modifications as their respective collocations render 
expedient. It was made by Mr. Dent, from plans sup- 
plied by Sir E. Beckett, then Mr, Denison, in 1854, and 
was fixed in the tower in 1859, since which time it 
has, by its pei-formance, given universal satisfaction and 
eflfectually silenced all adverse criticism, its variation, as 
certified by the then Astronomer Royal, Sir G. Airy, in 
1872, averaging one second per week. 

The French have a few turret clock factories, and 
have turned out some very elaborate and some very 
good clocks, but they are considerably dearer than ours. 

The only other class of clocks now made in this 
country are the eight-day dials. These clocks, made in 
various sizes, are used for offices, kitchens, schools, 
halls, &c., and are good serviceable clocks, where not 
much ornament is requii*ed; they are made both with 
and without striking parts, and have the great advantage 
over all foreign spring clocks of having the adjustment 
of the fusee, which, in all spring clocks with short pendu- 
lums, is essential to compensate for the varying force of 
the spring, and without which it is impossible to get 
anything like good time from them. 

Nearly the whole of the rest of the market is in the 
hands of the French and Americans ; the French com- 



manding it for the marble, gilt, and carriage clocks, 
while the Americans have nearly their own way with 
the cheapest kitchen and small bed-room timepieces and 
alarums, which have almost entirely displaced the German 
and Dutch ones. 

Since the development of electricity as a motor it 
has from time to time been applied to clocks, but hitherto 
with no very successful or practical results, though its 
further development may possibly lead to a system of 
controlled clocks, of which we, and London especially, 
are greatly in need. 

The introduction of a recognised public standard of 
time, either by controlled clocks or by hourly time 
signals, would be a great boon both to the public and to 
the makers of high-class watches. It is a want which 
has long been felt, and which watchmakers have long 
desired to have supplied. 



13. The Earliest Watches. — The first watches were 
made at Nuremberg, about the year 1500, where the 
mainspring is said to have been invented by Peter Hele, 
a clockmaker of that town ; they were furnished with 
the verge escapement, similar to the one in De Wyck's 
clock, the want of control over which no doubt sug- 
gested the idea of the fusee adjustment, which was 
invented about the year 1520. Charles V. of Germany 
is said to have taken great interest in the performance of 
these early specimens of horological art, and to have kept 
several of them going together. They appear to have 
been in common use in England in the reign of Queen 
Elizabeth, and are mentioned by Shakspere in " Twelfth 
Night." But it was not until the invention of the balance 


spring, about the year 1660, that any rapid progress was 
made in the direction of good timekeeping. 

14. Hooke*S Law. — The invention of the balance 
spring is assigned both to Huygens and to Dr. Robert 
Hooke ; biit, notwithstanding the genius of the Dutch 
clockmaker and philosopher, the merit of its applica- 
tion must rest with Hooke, whose enunciation of the 
theory of its isochronism in the words * Ut tensio sic 
vis " — " As the tension so is the force " — showed that 
he was well acquainted with its properties, although 
this axiom, which is called "Hooke's law," is true only 
of certain lengths of spring fixed at the ends at certain 
positions. This will be considered more fully in the 
chapter on the Balance Spring. 

15. The Clockmakers' Company. — The Clockmakers* 
Company of London was incorporated under a charter 
granted by Charles L, on the 22nd of August, 1631 ; 
and David Ramsay, clockmaker to the king, was 
elected the first master; but they did not get their 
livery until 1767. They had the power of regulating 
the watch and clock trade in, and for ten miles 
round, London, and the right of entry into vessels, 
shops, and warehouses for that purpose. They were a 
very active and representative body, as their recoi-ds 
will show — entries being not uncommon which state 
the fact of ."bad" or "deceitful" watches, etc., being 
broken. up — until the present century, when, Free Trade 
obtaining, they relapsed into that state into which 
bodies, corporate or otherwise, generally fall that have 
nothing to do. However, since the formation of the City 
and Guilds of London Institute for the Advancement 
of Technical Education, most of the Companies have seen 
the desirability of identifying themselves with the trades 
from which they take their names, and the Clockmakers' 
Company now ofier the freedom of the Company and 
a prize of ,£10 to the most successful competitor at the 
annual Greenwich trials, and a prize of £5 to the second, 
subject to certain conditions; they also grant .£10 per 


annum to the Korological Institute for the purpose of 
educating watchmakers. But since any one who can buy 
a chronometer (that may happen to be first) and have it 
in his possession for a certain length of time, can obtain 
this honour and reward without being able to make any 
part of the instrument, this action of the Company is 
hardly sufficient to promote technical education or other- 
wise to materially assist the' trade. Unfortunately 
they have not a hall of their own, and some years 
since they handed over the valuable technical library 
and objects of trade interest they possessed to the cus- 
todians of the Guildhall Library and Museum, an insti- 
tution which, though of the utmost utility to the general 
public, is almost inaccessible to the great majority of 
watch and clock makers. 

16. Bepeating work was invented in the year 1676, 
by a clergyman named Barlow. He got Thomas Tompion, 
the celebrated watchmaker of that period, who had as- 
sisted Dr. Hooke in his researches, to make a watch for 
him, with his repeating mechanism attached, and applied 
for a patent for it ; but he was opposed in his applica- 
tion by Daniel Quare, a watchmaker, who, some years 
later (in 1691), introduced the minute hand, and who 
had been devising something in the same direction. The 
matter was referred to the king (James II.), who, upon 
tiying the pieces submitted to him for his decision, gave 
his judgment in favour of Quare. This striking work 
was subsequently much improved upon by Julien Le 
Roy, of Paris, who introduced a piece which prevented 
the repeating part from striking until the rack was 
driven home to the snail or stop, and consequently 
insured its striking the hour correctly. The English 
repeaters were made to strike on a bell screwed into the 
back of the case and surrounding the movement, and it 
is to the Swiss, later on, that we owe the substitution of 
springs upon which the hammers strike. 

Although the repeating watch was an English inven- 
tion, and the motion work made by Stogden was a further 



improvement of it, the making of the repeating motion, 
or striking parts, of watches has been entirely abandoned 
in this country ; the branch is a long one, the work is 
entirely done by hand, and is not very remunerative, and 
there is not a sufficient number of repeaters made to em- 
ploy many hands, so that the Swiss have it all to them- 
selves, and English watchmakers are obliged of necessity 
to obtain this part of the watch from Switzerland, though 
the best repeaters are made up here. 

17. Jewelling, and the Horizontal Escapement. — 

From the latter part of the seventeenth century, the 
science of horology steadily improved in this country. 
Facio introduced jewelling from Greneva in the year 
1700; and George Graham, an apprentice of Tompion, 
invented the cylinder escapement in the same year. This 
escapement, although found to be a great advance on the 
verge, was soon almost discarded, owing to the rapidity 
with which the brass escape wheel cut the cylinder, 
which was of steel. This indeed caused the introduction 
of ruby cylinders (some of which, made by Mudge and 
his contemporaries, are still the admiration of our present 
experts, and proofs of the manipulative skill and patience 
of our early horologists), but the difficulty and expense 
of making rendered them unsuitable for ordinary watches. 
The Swiss, however, took the idea up, and the horizontal 
escapement with a small steel escape wheel was, until 
very recently, almost the only one applied to their 
watches ; it is indeed generally made now, but the supe- 
rior qualities of the lever are slowly asserting themselves 
in that country, and are gradually driving it out of the 

18. Harrison's Marine Timekeeper. — The want of a 
means of discovering the longitude at sea with any 
degree of certainty having long been felt amongst 
maritime nations, and by England especially, which 
was at that period rapidly developing her commercial 
activity, an Act of Parliament was passed in the year 
1714, the twelfth year of the reign of Queen Anne, con- 


stituting certain persons commissioners for the discovery 
of the longitude at sea (called afterwards the Board of 
Longitude), and offering the very large sum of £20,000 for 
the best means of attaining this object. This Board evi- 
dently thought a good timekeeper what was needed, and 
offered sums of respectively^! 0,000, £15,000, and£20,000 
for one that would determine longitude to within sixty, 
forty, or thirty mUes. There is no authentic record of 
any horologist attempting to solve this problem, and gain 
any of these rewards, until the year 1728, fourteen years 
after they were offered, when John Harrison appeared in 
London, and submitted plans and drawings of a marine 
timekeeper te Dr. Halley, then Astronomer Boyal. The 
Docter recommended him to George Graham, who saw 
the merit of his plans, but, being a practical man, ad- 
vised him te go home again and make his machine before 
applying to the Board for any assistance. 

John Harrison was the son of a carpenter, and was bom 
at Faulby, near Pontefract, Yorkshire, in the year 1693. 
His father had brought him up to his own business, but 
he was early devoted to mechanical pursuits, and had 
spent much of his time in repairing clocks and watehes. 
He had experimented on pendulums, and had invented 
the gridiron compensation, this invention giving him a 
position in the eyes of Graham, who was himself occupied 
on the same subject ; and his experiments on the effects 
of temperature on various metals no doubt suggested te 
him the application of the principle of compensation to 
watehes, in order to counterbalance the variations in rate 
caused by the expansion and contraction of the balance- 
spring in heat and cold. 

Harrison's compensation was effected by a laminated 
piece fixed to the plate at one end, and carrying curb 
pins, between which the spring acted, at the other ; and 
it was by the movement of these pins to or from the 
stud into which the spiing was pinned, lengthening or 
shortening the acting portion of the spring, that the 
wateh was regulated, this curb acting in the same manner 


as a watch is regulated by hand. But notwithstanding 
Harrison's use of the brass and steel pieces pinned 
together, his suggestion that the compensation should be 
eff'ected by the balance, and his acquaintance with Graham 
and other famous English watchmakers, the honour of 
constructing the first compensation balance is acknow- 
ledged to belong to Julien Le Roy, the famous French 
watchmaker, who also invented a mercurial compensation 

Le E/Oy's balance was made in the same manner as 
Hari'ison's " thermometer kerbe," as he called it, namely, 
by pinning the laminsB of brass and steel, composing 
the circumference, together. The first Arnold's balances 
were also pinned in this manner, and Reid speaks of 
them as performing better than those that were after- 
wards made solid, when Thomas Eamshaw had made an 
improvement, amounting to an invention, by forming the 
rim out of a solid piece by melting or fusing the brass on 
to the steel. 

The escapement of Harrison's timepiece was a sort of 
remontoir, very complicated and very beautiful, as may 
be seen in the fourth chronometer of his construction, 
now in the Royal Observatory, Greenwich, a full descrip- 
tion of which is given in Vol. I. of the Horological 
Journal, In the year 1735 he again came to London, 
having in the seven years' interval constructed his time- 
keeper ; and it seems curious that, although nearly a hun- 
dred years had elapsed since the first application of the 
spiral spring by Hooke, there is no record of any at- 
tempt having been made to counteract its irregularities 
consequent upon variations of temperature (and which 
rendered timekeepers of comparatively little use for 
astronomical and other scientific purposes) until this 

Harrison, having satisfied himself of the satisfactory 
performance of his timepiece by trials on board a barge 
on the Humber, obtained permission in the following 
year to proceed to Lisbon in one of the king's ships. In 


this voyapfe he was able to correct the reckoning to 
within 1" 30', and the Board of Longitude granted him 
X500 to still further improve his timekeeper. 

After constructing other chronometers with various 
improvements, he was enabled to ascertain the longitutle 
at sea to within ten miles, one-third of the distance 
required by the Act of 1714; but it was not till the 
year 1757, after much trouble and many remonstrances 
with the Commissioners, that he obtained the last portion 
of the ,£20,000 offered by Parliament, only nine years 
before his death, which took place at his house in Red 
Lion Square. He was buried in Hampstead churchyard, 
and his tomb, which had fallen into decay, was restored 
a short time ago by the Clockmakers' Company. ' 

Harrison must have been a man of extraordinary 
genius and perseverance, having had so many disap- 
pointments to encounter. He invented the maintaining 
power which rendered practicable the use of the fusee in 
spring clocks and watches, without which marine chrono- 
meters could never have been brought to their present 
perfection. It is the most perfect adjustment of the un- 
equal pull of the mainspring, and has been the great and 
distinctive feature of the English watch up to the present 

The third chronometer of his make, together with a 
facsimile of it by his apprentice, Mr. Kendall, is also 
preserved at Greenwich Observatory. It is in a very 
large double silver case, shaped like an old-fashioned 
watch, and is of very beautiful workmanship. 

19. Mudge's Chronometer. — After Harrison had re- 
ceived the last portion of the reward, another Act of 
Parliament was passed, limiting any further reward to 
.£10,000, and also prescribing a much closer rate of 
going for timekeepers, it being taken for granted that 
as the attention of so many eminent men had now 
been directed to this importaiit subject, a better instru- 
ment than Harnson's would soon be produced. The 
Astronomer Royal, Dr. Maskelyne, being now appointed 


to test the instruments, Thomas Mudge (a watch- 
maker who had long been preparing a timekeeper 
to compete with those of Harrison) submitted one for 
trial, which, after a twelve months' test, although it did 
not come within the limits prescribed by the Act, was 
reported so favourably of that the Board of Longitude 
granted him £500. Mudge mside other and improved 
chronometers, he, too, having trouble with the Commis- 
sioners. But after an inquiry into the merits of his 
inventions, and in opposition to the Board of Longitude, 
the House of Commons made him a further grant of 
£2y500. He, in 1750, invented the lever escapement, 
which is, with some modifications, the same in principle 
as that now in use, and is acknowledged to be the best for 
pocket watches at the present day ; but, owing possibly 
to the difficulty of getting it well made, or to its wrong 
proportions, its manufacture was abandoned in favour of 
the verge and horizontal escapements, and even of the 
rack lever, an inferior escapement on a similar principle, 
patented by Peter Litherland, a Liverpool manufacturer, 
in 1794 ; and it was not until George Savage improved 
and altered it that the true value of the lever escapement 
was appreciated. 

20. Arnold's Escapement. — John Arnold had, at the 
same time as Mudge, a chronometer under trial at Green- 
wich, which was reported to perform better than Mudge's, 
and although Arnold's chronometer was not undergoing 
the trials with a view to a Government reward (as neither 
he nor his son ever placed a chronometer there with that 
idea), his instrument attracted the attention of the com- 
mittee whose business it was to inquire into the merits of 
these timekeepers and the principle of their construction, 
the result being that, from the very favourable report of 
it, Mr. Arnold received at diflferent times a sum of 
j£l,322 from the Board of Longitude, and his son, in 
1805, received the remaining portion of a grant of 
£3,000, an amount equal to that granted to Mudge. 

Arnold, in 1782, patented the detached or " chrono- 


meter" escapement and the compensation balance, and 
although both these inventions are claimed by the 
French for Julien Le Roy, Le Roy*s idea (at all events of 
the escapement) must have been very crude, and could 
scarcely have amounted to much more than a suggestion. 
But as the French Government had also at this time 
offered large rewards for a perfect timekeeper, it is very 
difficult to say how much of any invention belonged to 
the person credited with it. However, the escapement 
patented by Arnold in 1782 was a long way in advance 
of anything that hsid preceded it, and was the parent, 
with but little modification in principle, of the chrono- 
meter of to-day. As has been mentioned, Harrison had 
advised the compensation of the spring to be effected by 
the balance, instead of by an application to the spring 
itself, and Julien Le Hoy had succeeded in makmg a 
compensation balance. Arnold's balance, which was a 
modification of Le Roy's, was made to compensate by 
affixing two laminated arms pinned together to the out- 
side of a circular solid rim ; he afterwards made the rim 
itself of these laminae. 

21. The Cylindrical Spring. — ^To Arnold also is due 
the credit of the invention of the helical, or cylindrical, 
spring, the "incurvating the ends" of which, he says, 
"is attended with the property of rendering all the 
vibrations of equal duration, because the figure is always 
similar to itself." There is no doubt that Dr. Hooke 
when he applied the spiral spring to the balance of a 
timekeeper perfectly understood its isochronous proper- 
ties, and his law of tension might have pointed it out 
to others; but, notwithstanding Harrison's undoubted 
genius, he could not have understood it, as he got his 
long and short arcs of vibration equal by the very objec- 
tionable method of weighting the balance more in one 
place than another, and, in explanation, he says "that 
the natural tendency of the vibration is to go slower in 
the short arcs." Arnold evidently knew better than this, 
and his detached escapement must soon have taught him 


the great importance of the isochronous properties of the 
spring, which, with its cylindrical form and curved ends, 
remains just as he left it ; therefore, although Harrison 
is justly remembered as the " father of English chrono- 
meter makers," the principles of Arnold's timekeepers 
have been the most useful and permanent. 

22. Earnshaw's Balance. — A name always asso- 
ciated with Arnold is that of his contemporary and rival, 
Thomas Eamshaw, who followed up his improvements. 
In the first place Eamshaw improved the balance by 
melting the brass on to a disc of steel in such a way that 
it could be turned down in the lathe to the proper propor- 
tions, and then become a solid piece of metal, or metals, 
instead of two pieces pinned together, or soldered together, 
as was sometimes done ; and this balance of Earnshaw's 
is now applied to nineteen out of every twenty marine 
chronometers without any alteration whatever.* 

But the improvements made by Eamshaw on Arnold's 
escapement were of even more importance than his im- 
provement of the balanca He was the last of what may 
be termed the chronometer makers of the eighteenth 
century, and his chronometer left so little room for 
improvements that no further considerable sums have 
been oflfered by Government to induce the makers to 
devote their time to that object. The annual trials of 
chronometers at the Royal Observatory, Greenwich, have 
been continued, and the Government purchase a few of 
the best of them, but, from being extremely liberal, they 
have become extremely economical, and pay such prices 
for those they do purchase as would not repay any work- 
man for the extra time and trouble necessary to be 
devoted to their preparation for such severe trials as 
they are subjected to, with the prospect of selling at very 

* Beid in his Treatise on Clock and Watch Making speaks of the 
balances made of separate laminae pinned together as performing 
better than those made after Earnshaw's method. It is, however, 
not eas^ to see how this could be, and ifc must have been exceed- 
ingly difficult to have made one of these balances anything like 


little more than the market price of the articla But, 
then, the title of "Chronometer Maker to the Admi- 
ralty" is worth money to many men who are not 
chronometer makers, while it is worth nothing to many 
who are, and who, therefore, refrain from troubling 
themselves to invent means of overcoming the effects of 
the absurd ranges of temperature through which the 
instruments are now tried, ranges through which it is 
utterly impossible they could be exposed in actual use. 

23. Auxiliary Compensation. — As the ordinary com- 
pensation balance has an inherent failing in its action, 
and only approximates to the action of the spring 
through a limited range of temperatures, or, if adjusted 
for extremes^ fails in the mean temperatures, it was 
found necessary, in order to counteract this error (which 
is genei-ally known by the somewhat misleading name of 
'* middle temperature error ''), to devise a secondary, or 
auxiliary compensation, as it is called, and to this end 
most of the efforts of chronometer makers have for many 
years been directed. 

The discovery of this error marks an epoteh in the 
history of chronometer making ; and the necessity of 
correcting it by means of an auxiliary compensation, 
mainly owing to the great range through which chrono- 
meters are tried at Greenwich — a range which some- 
times extends over more than 100" F. — has brought out 
the inventive faculties of chronometer makers and others 
in a truly astonishing manner. The late Mr. Charles 
Frodsham, speaking of horological inventions, said that 
"if every inventor would record his failures it would be 
a boon to those who might be inclined to follow in his 
footsteps," but most of the so-called improvements are so 
hopelessly foolish that I will only enumerate a few of 
the balances that have been constructed on somewhat 
sound principles by some of our best men down to the 
present time. Professor Hartnup of the Liverpool Obser- 
vatory (who, from his great experience of chronometers 
and their rates, may be considered one of our highest 


authorities on the subject), in laying down the principles 
of their construction, has given the following rule : — 
"The balance-rim must be of a circular form, so that 
the laminse of brass and steel may be turned down to 
the requisite proportions with facility, and the compen- 
sation and poising easily effected,'' and, I wiD add, surely 
maintained, as no timekeeper can perform accurately for 
any length of time if the balance is of such a form as to 
get out of shape, and consequently out of poise, by any 
variation of temperature, or by the rough treatment 
chronometers are almost sure to be subjected to on board 
ship. Chronometer balances are often found to have 
taken what is called a " set " after having been some 
time at sea (the rim having become distorted), and the 
chronometers frequently in consequence have changed 
their i*ates. This must be owing to their being exposed 
to either greater heat or greater cold than was the case 
when their adjustment was effected. But, if the balance 
be well made and true, and the laminse of equal thick- 
ness throughout, the limbs will assume the same form, 
and it will seldom be found much out of poise, and should 
the compensation not be altered by this alteration of the 
circle of the balance, it is unwise to interfere with it 
further than to see that the poise is not affected. 

The successful employment of the simple and service- 
able form of balance bequeathed to us by Eamshaw, 
which is adopted in nearly all ships' chronometers, is 
mainly owing to its near approach to the conditions laid 
down by Professor Hartnup. Numerous other contri- 
vances (invented to overcome the error of the balance), 
some of them of the most grotesque and unseemly shapes, 
are mostly to be found in horological museums, or in 
the lumber drawers of chronometer springers, who have 
had to remove them when they were found not to 
answer. Many of these have been taken from chrono- 
meters which gave good results in the Greenwich trials, 
but were discovered to be quite useless for the every-day 
work on board ship. 


24. Improved Chronometer Balances. — The first 
record we have of any attempt being made to over- 
come the "middle temperature error" is that of a 
chronometer maker named Hardy, who in 1804 received 
a prize of £30 from the Society of Arts for an improved 
chronometer balance (see p. 244). Hardy rejected the 
old circular form, and adopted that shown in Fig. 47. 
There appears to be no record of its performance, but 
it was the forerunner of others of the same form, or 
a modification of it ; the original Mr. Dent evidently 
took the idea of his balance from it, and improved upon 
it (see p. 247). Other chronometer makers were not 
idle during the years following Hardy's invention. Moli- 
neux patented a balance with an auxiliary compensation, 
which the late Mr. Eiffe claimed as his invention, 
and over which a good deal of discussion and bitterness 
resulted. Some few months before his death Mr. Eiffe 
wrote me a letter in reference to something I had said 
at the Horological Institute respecting this balance, in 
which he said his claim to its invention had been acknow- 
ledged by the Admiralty, who had given him the sum of 
j£300, on the recommendation of the Astronomer Royal, 
in consideration of its merits. Chronometers of both 
Eiffels and Molineux's construction with this improve- 
ment were tested at Greenwich, and were reported to 
have given good results ; but the balance in its then form 
has never, for various reasons, come into use. 

Loseby, a celebrated chronometer maker, about the 
same time patented a modification of the mercurial com- 
pensation invented by Julien Le Roy, and with his 
skilful manipulation this balance gave wonderfully good 
results for a series of years at the Greenwich trials. He 
applied several times for a reward, sending in a list of 
the rates of going of his and other chronometers, by 
which he endeavoured to prove that his were the most 
successful in trials extending over a period of five years. 
This produced a denial from the first Mr. Dent, who 
claimed the superiority for his own chronometers, and 


the dificussion was afterwards continued by Sir Edmund 
Beckett, who, with his usual acumen, had considerably 
the best of the controversy. Loseby, disappointed at not 
getting a reward from Government, withdrew altogether 
from the trials. 

Poole's application to the ordinary balance may in 
point of time next be noticed. It is of the simplest con- 
struction, and, perhaps in consequence, may be reckoned 
among the most successful It has many times enabled 
chronometers fitted with it to take high places at Green- 
wich, and has probably been more generally adopted in 
commercial chronometers than any other form of auxiliary 
adjustment, if it can be called by that name, as its only 
action is to oppose a slight resistance to the rim of the 
balance when moving outwards in cold (see p. 250). 

25. Hartnup's Balance. — In the many opportunities 
he has had of testing the chronometers sent for trial at 
the Liverpool Observatory, Professor Hartnup has had 
ample scope for thoroughly investigating the so-called 
middle temperature error, and the necessities for, and 
requirement of, an auxiliary compensation ; and he has 
invented a balance which, while securing the continuity 
of action of those of Hardy and Dent, kept within the con- 
ditions laid down by himself and already referred to. 

Mr. KuUberg modified Hartnup's balance, and pro- 
duced what is known as Kullberg's flat-riinmed balance 
(see p. 253). Chronometers with this balance have stood 
first at Greenwich on several occasions, but it has not been 
adopted by the trade, partly on account of the difficulty 
of making and adjusting it, but principally owing to its 
want of rigidity and strength to withstand rough usage. 

26. Mercer's Balance. — Another compensation that 
has given the best result and is about the best auxiliary 
I have seen or tried is a modification of the invention 
of Molineux, or Eiffe (or both as it may possibly have 
been), due to Mr. Mercer (see p. 256), who, with this 
balance, has won many Admiralty prizes both for him- 
self and others. I have called it a modification, because 


the principle of action is the same as in Molineux's, but 
it is more of an invention than many alterations that are 
called inventions. There is no di&culty in making it, 
there should be little or no diflficulty in applying it to any 
ordinary balance, it is easily adjusted, and is very little 
more liable to disarrangement than the ordinaiy balance. 
The auxiliary pieces are fixed to the arms of the balance 
with a steady pin and two screws, and the soles or platform 
which support them are perfectly solid and allow them to 
act with as much regularity and precision as the limbs of 
the balance itself. As the error of the balance increases in 
an increasing ratio with increments of temperature, this 
auxiliary acts only at high temperatures, from whatever 
point may be determined upon in adjusting it, the adjust- 
ment being regulated by two screws tixed in the rim of the 
balance coming into contact with the laminated arms of 
the auxiliary and stopping their action at any iixed 

As the regulating power of the balance is dependent 
on the distance of the moment of inertia from the centre, 
and as this is proportional to the square of the radius 
of the mass from that centre, it foUows that the farther 
out the timing nuts of a balance are drawn, the greater 
will be the error of the balance through their expansion 
and contraction from and to the centre, not only from the 
variation in the length of the balance arm, but from their 
own variation and that of the small screws which carry 

This error is of so much importance that it should not 
be lost sight of, and in no case should these weights or 
screws be drawn outwards beyond the circumference of the 
compensation weights ; and if, when the adjustment of the 
chronometer is complete, it is found necessary to draw 
them out, the two smaller screws at the side of them, 
which are usually added, should be replaced by heavier 
ones, and the weights screwed home to within one turn or 
thereabouts ; indeed it would be a further improvement 
if they were placed inside the rim, as then any expansion 


that might take place in them or their screws would be 
carrying them towards the centre instead of away from it, 
and would assist in diminishing the error against which 
all these complications are directed. 

The adjustment of marine chronometers is now effected 
by means of two weights sliding on the rim of the balance. 
Formerly screws, such as are still necessarily used for this 
purpose in pocket watches, were sometimes used ; but it 
has been found that, where there is room, weights are pre- 
ferable, as the laminae composing the rim are kept solid, 
drilling and tapping holes all round it being certainly a 
great evil. Some of the old pocket chronometers have 
adjusting weights on the balances, but there is no room in 
watches of modem size and shape for them. 

27. Compensated Watches. — No good watch is now 
made without a compensation balance, but there are also 
many bad ones made with them, and it has lately become 
the custom to put what are called compensation balances 
to the very woi*st of the foreign watches sold in this 
country ; these balances being infinitely worse than brass 
or steel ones, especially if they are cut open, as the mate- 
rial of which they are composed is so soft that the least 
touch puts them out of shape, and consequently out of poise, 
so essential to the going of even the worst of watches. 

Cheap watches with compensation balances should be 
discouraged by both the sellers and wearers of them, and 
instead of a compensation balance in such a watch being 
a recommendation it should be considered as indicating 
a sham, which it generally is. 

Nine-tenths of the English watches that have compen- 
sation balances have never been adjusted, and although 
a watch with a well-made compensation balance will 
generally give better results than one with a solid metal 
one, there is no rule by which it can be made approxi- 
mately near, and it is only by the usual testing in the 
oven and in cold that the performance of any balance can 
be tested ; but as the principal extra cost of compen- 
sated watches is not in the balance but in the somewhat 


tedious and sometimes troublesome process of adjusting 
it, this last operation is omitted, and the public (not five 
per cent, of whom know anything of the matter) buy them 
and find out their errors afterwards. Even if the manu- 
facturer is honest enough (and he usually is) to tell the 
truth in the matter, the practice of making these watches 
is a bad one, as may easily be seen if the fact I have just 
stated — and I think I have overstated the proportion of 
adjusted watches made — be compared with the advertise- 
ments everywhere to be found of watches " perfectly 
compensated and adjusted for all temperatures," to be had 
at very moderate prices. 

As the development of the compensation has been 
identical with the development of watches and chrono- 
meters, it has been necessary for me to trace it from the 
earliest efforts down to the present day. 

28. Ships' Chronometers. — If the spring of a watch 
or chronometer be perfectly isochronous, the balance 
should perform tbe Vibrations in the same time, irre- 
spective of the number of degrees passed through, but 
as, in practice, there is no such thing as perfect iso- 
chronism (although watch-springers, by dint of very 
careful manipulation, may make their timekeepers ap- 
proximate very nearly to that state), ships' chronometers 
are now always hung suspended in gimbals, in order that 
by thus maintaining the balance pivots in a vertical posi- 
tion, the side friction on them (which is much greater 
than the end friction and would consequently greatly 
diminish the arc) may be avoided and the arc kept more 
constant. Of course this is only a help towards good time- 
keeping, as the thickening of the oil, which is sure to take 
place after the chronometer has been going for any length 
of time, will have the effect of shortening it. 

To still further promote the good performance of the 
chronometer, the box in which it is suspended is placed 
within a thickly-padded outer case, which maintains it at 
a more uniform temperature. 

29. English Watches. — The head-quarters of the 



English watch and chronometer trade is GlerkenweU, 
the remainder of the watches (mostly of the commoner 
kind) being chiefly made in Liverpool, Coventry, and 
Birmingham. Since the demand for cheaper watdbes on 
a large scale, consequent on the large number of foreign 
importations, watch factories have been established at 
the two latter towns. 

But hitherto the attention of English watchmakers 
has been directed to the improvement in the quality of 
watches, rather than to their cheapness of production, and 
it is greatly to be questioned whether the introduction of 
the factory system, and the wholesale adoption of machi- 
nery in the manufacture, would at all benefit the trade in 
this country. English watches have always commanded 
their price in the markets of the world, and the good name 
they have ever borne has led to their imitation by foreign 
producers, both by the forgery of English makers' names 
and of the English hall-marks in the cases, and by send- 
ing unfinished cases here and getting them marked at the 
Goldsmiths' and the provincial halls through the agency 
of case-makers and othera. This hall-marking, though in 
itself no trade mark, but merely a guarantee of the quality 
of the metal, is found useful by foreign makers in helping 
them to pass off their watches as of English make. This 
is proof, if any were required, of the high value set upon 
English-made watches, and of the estimation in which 
they are held by foreigners. And what would become of 
this esteem if, instead of continuing to maintain our high 
position, we were to compete with the slop trade of Swit- 
zerland or America ? There are many people who will 
tell us that some wretched thing in the form of a watch 
" goes well enough for them," but is such an argument a 
sufficient reason for lowering our standard of excellence ] 
No 1 If we are to make a more marketable article in the 
way of watches let us proceed in the right direction, by 
organising our system of manufacture, by altering and 
improving the plan of our movement (not by the elimina- 
tion of the fusee, but by, amongst other things, the total 


abolition of the full plate in favour of three-quarter- and 
half-plate movements), and by imparting more intelligence 
and education among our workmen. 

The want of technical education has kept the trade 
back more than anything else. Under the old appren- 
ticeship system, the pupil was usually kept running 
eri-ands and making himself generally useful for the 
fir^t year, or perhaps more, after which he was pro- 
moted to the bench, when, after learning perhaps to 
make pivots, a little filing, (fee, he was very often left 
to his own resources to pick up whatever he could. — 
(An amusing anecdote was once told me by a work- 
man of his early training as an apprentica The pro- 
cess of his instruction being that if the piece of work 
was satisfactory which he had completed, well and good ; 
nothing was said; but if not he was incontinently 
"knocked off the stool" by his gentle master, but the 
work, nevertheless, " went into the box.") It was with 
the empirical knowledge thus obtained that most of the 
past generation of finishers and others were formed 
into workmen, and it is not astonishing that the watch- 
manufacture of the country has not taken a higher 
rank amongst the sciences, when it is considered in 
whose hands it has hitherto been left ; indeed, the 
surprise should be at the progress made by men who 
have all along been working by what may be termed the 
rule of thumb. 

30. Apprentices and Pupils. — The apprenticeship 
system, with its seven long years of probation, although 
it still has its advocates, may be considered to all intents 
and purposes dead ; and perhaps it is as well that it is so. 
It is not at all suitable to modern ideas or tastes, and, 
even its present advocates, who are mostly men who 
have themselves endured it, whilst praising it in one 
breath, condemn it in the other, by citing their own ex- 
periences. Its place will no doubt be taken by schools 
of instruction, such as that of the Horological Institute, 
where, it is safe to assert, a pupil who is at all diligent 


Fig. 5 represents the separate parts of the fusee. 
A is the bottom of the fusee cone, which is hollowed 
out to receive the ratchet wheel, shown in position. 
This ratchet wheel and the fusee aAor are formed in one 
piece in three-quarter plate watches, and are fixed to the 
cone with three screws, the heads of which are sunk in 
the wheel ; the arbor goes right through the cone and 
terminates in the winding square (shown at f) ; the 
ratchet wheel is nearly flush with the outer edge of the 
fusee brass. B is a thin steel wheel with clicks and 
springs projecting from its surface that act into the teeth 
of the ratchet wheel A. The use of these clicks is to 
permit the fusee to turn one way when the chain is 
wound on to it, and to prevent it from turning the other 
way without the steel and great wheels when the chain 
pulls it in that direction ; the ratchet teeth of this wheel 
are cut in the contrary direction to those on the wheel at 
A. D and E are back and front views of the great wheel ; 
E shows the maintaining-power spring let into a groove 
in the wheel, one end of which is fixed to the wheel by 
a pin going through, while the other end is free to move 
the distance of the slot ; the pin in the free end of the 
spring projects, one end into the slot shown in D and e, 
and the other end into the hole in the ratchet wheel B, 
thus preventing these two wheels from moving more 
than the distance of the length of the slot independent 
of one another. Both the wheels are fitted to move freely 
on the fusee arbor and are kept in their plaoes by the 
collet shown at c, 1 and 2, in plan and elevation, which 
is fixed by a pin which passes through its pipe and the 
fusee arbor. A click, called the fusee detent, pivoted 
into the frame of the watch is kept in contact with the 
teeth of the wheel B, by a weak spring screwed to the 
pillar plate ; when the force of the mainspring (which 
must always be stronger than the maintaining-power 
spring) is exerted and draws the fusee in the direction of 
the barrel, it will carry the steel wheel and the end of 
the maintaining-power spring to the end of the slot 



in the great wheel ; the great wheel cannot move, from 
its teeth engaging with the centre pinion, and the steel 
wheel is held where the mainspring has drawn it by the 
click or detent. If the power of the mainspring is taken 
off by winding, the maintaining-power spring will exert 
sufficient force on the great wheel to keep the watch or 
chronometer going for a few minutes (more or less, 
according to the length of the slot in which the pin 
at the end of it acts). F is the upper end of the fusee, 
showing a steel cap fixed to the brass cone with two 
screws, the projecting hook of which stops the winding of 
the watch at the proper time, namely, when the chain is 
all wound round the fusee. The barrel arbor has a square 
cut on it, to which is fitted a small ratchet wheel on the 
dial side of the pillar plate, and a click (screwed to the 
plate fitting into the teeth of this wheel) prevents the 
arbor and wheel from turning when it is screwed down ; 
this click and ratchet wheel are used only for setting the 
spring up sufficiently to get an adjustment of the main- 
spring ; when the spring is adjusted, the click is screwed 
down, and the barrel arbor remains stationary. 

The fusee (Figs. 5, 6) has four-and a-quarter turns on 
it, which allows of the barrel turning three times on its 
arbor, and with a great wheel of 84 and a pinion of 12 the 
watch will go for thirty hours, giving six hours' grace in the 
case of irregular winding. The large end of the fusee should 
have a diameter double that of the smaller, and the cone 
should be slightly concave. There are some very com- 
plicated formulsB for calculating the proper shape of the 
fusee, but as the form must necessarily alter with the 
number of turns of the spiral cut upon it, and as the 
mainspring may be more or less taper, no rule can be 
stated for this that would serve any useful purpose ; a 
fusee of the form given will, however, be adjustable. 
The larger end of the fusee and the barrel should be of 
the same diameter as the rim of the great wheel, the 
teeth projecting from the fusee about as much as does 
the chain from the barrel, and both should be as large as 


possible, as the larger they are in diameter the greater 
will be the power in proportion to the size of the watch. 

46. Fusee Stop. —The usual stop to a fusee watch is 
a straight piece of steel fitted to a stud in the top plate, 
placed so that the end of it will just catch the projecting 
hook of the fusee-cap when in the same plane ; it is kept 
free of the hook by a weak spring underneath it until the 
chain comes on the last turn of the fusee, when it presses 
the stop close to the plate, and the hook catches it. 

This stop is a very good one, where there is room for 
it and it is well and carefully made, but in common 
watches it is seldom well made, and in very flat and full- 
plate watches there is not sufficient room for it, and its 
failure to act is often the cause of cliains breaking. A 
much better stop is the solid one which is made in one 
piece, since instead of the blade rising at an angle, as the 
stop described does, it is raised and lowered parallel to 
the plate, thus giving more freedom to the fusee hook to 
pass, and, as there is no stud required, it occupies less 
room. It is fixed to the plate with two screws, the space 
between the screws forming the spring ; the screw at the 
bend has a shoulder, and the hole in the blade fits 
loosely on the plain part of the screw, the head of the 
screw being just far enough from the plate to allow the 
stop to rise the required height to free the fusee hook ; 
the spring may be made very thin, as there is no pressure 
on it. The end of the stop should be flat, or sloped a little 
outwards, so that with any undue strain the hook would 
press it against the plate, and it should be planted 
opposite the fusee arbor, the blade forming a right angle 
with the face of the fusee hook when in contact with it. 
If the blade is left too long, as it often is, the hook 
pushes it away, and if too short, pulls it towards the centre 
of the fusee, thus frequently passing it and breaking the 
chain. This stop is seldom applied, and never (where it 
is most wanted) to full-plate watches. It certainly 
ought to be more easily made^ and therefore cheaper, 
than the common one if it were universally adopted, as it 


could be stamped out of sheet steel the required shape 
and thickness. 

There is another very good fusee stop^ formed of a 
straight piece of steel filed thin to act as a spring, with 
a step or hook near the outer end that catches the fusee 
hook when the chain, acting on the extreme end, presses 
it down. This stop has a great advantage over the 
others, inasmuch as, instead of the fusee hook pressing 
on the end and pushing it towards its attachment to the 
plate, it catches the stop or hook and pulls it from its 
attachment ; but it can only be employed when the chain 
is acting on the inside of the fusee, with a different 
arrangement of the calibre, commonly called a left- 
handed movement, i.e., a movement where the usual 
positions of the barrel and fusee are reversed. It has 
been pointed out by writers on this subject (as I have 
before remarked, in treating of Movements), from Mudge 
to Sir K Beckett, that the usual arrangement in English 
watches is wrong, and the Swiss evidently acted upon 
tliis knowledge, as most of their pocket chronometers 
(and pocket chronometers are the only Swiss watches 
possessing a fusee) have the chain on the inside of the 
fusea If the chain acts on the opposite side of the 
fusee to the centre pinion, the pressure on the fusee pivots 
is the sum of the force of the mainspring on the fusee 
and of the great wheel on the pinion ; but if the chain 
acts on the same side as the pinion it will only be the 
difference. The friction on the fusee pivots would there- 
fore be greater in the former case than in the latter, and 
it is well known that the wear is greater in the fusee 
holes than in those of the rest of the train. In reference 
to this. Sir E. Beckett says : — " I know no reason why the 
common arrangement should be adhered to, except that 
it is the common one, which is generally considered 
reason enough for anything bad." There is much force, 
and, when applied to watchmaking, much truth in this, 
otherwise the full-plate watch would not have kept its 
supreme position until it has, with other matters of the 


same sort, greatly injured the English watch trade. 
There are reasons, however, why this caliber has not 
been adopted. The finisher had acquired from practice 
great facility in running in the fusee and barrel in the 
usual way ; making and adjusting the stop is more 
difficult when the chain is acting inside than when out- 
side the fusee, and manufacturers did not see sufficient 
practical advantage in the change to make them pay 
more for it. The top fusee and bottom centre wheel 
holes and pivots are more liable to wear than any other 
parts of the watch ; but the fusee holes are generally left 
too short, and dust or damp is introduced by the key or 
through the keyhole in the case ; the bottom centre hole 
is likewise nearly always too short, and the pivot close to 
the pressure of the great wheel ; the cannon pinion also 
is usually fitted so close to the plate that it draws the oil 
from the centre hole. But although the centre holes are 
well known to be subject to wear, they are seldom 
jewelled, even in the best watches ; no good watch should 
be made without jewelled centre holes. 

47. Mainsprings. — The employment of the mainspring 
as a prime mover for clocks and watches, although dating 
from the sixteenth century, has undergone little or no 
change in the manner of its application ; its action is 
unique. I know of no other mechanical combination 
wherein a spring is required to maintain such sustained 
and uniform resistance for so long a period, nor in 
which it is expected to last so long. 

As the manner in which the power is given out from 
the spring is second only in importance, in the construction 
of a timekeeper, to the controlling and regulating of the 
vibrations of the balance by the balance spring, the 
action of the mainspring requires, and has received, a 
good deal of consideration from watchmakers, especially 
those of France and Switzerland. In England also, when 
watches were made with verge escapements, the adjust- 
ment of the mainspring was considered of paramount 
importance, as the time of the watch varied so much 


with that escapement with the lea^ irregularity in the 
motive force, but when the lever escapement became 
universal, the English watchmaker troubled himself very 
little about adjusting the mainspring. 

When the number of revolutions the great wheel was 
to make was determined (and they were generally nearly 
the same), every one concerned knew the form of fusee 
that would give an approximate adjustment with the aid 
of the barrel ratchet, and there was never any question 
or discussion as to the best manner of attaching the 
spring to the barrel. Every finisher knew how to hook in 
a spring in the best possible manner (and could generally 
do it, as it was the first piece of work he was taught to 
do as an apprentice), that is, a square hook on the spring 
fitting a corresponding hole in the edge of the barret In 
common watches these hooks were frequently made too 
large, which necessitated a large hole in the barrel, and 
weakened it very much at that part, and as these 
watches were gilded for show, and made soft so as to get a 
surface on the gilding, if a strong spring broke in such 
a barrel it would bulge it out at this weakest part, when 
either a new cover would have to be made or the old one 
be hammered out. This weakening of the barrel has been 
urged as a reason for adopting the foreign method of 
attaching the spring to the barrel by means of a pin in 
the side of the barrel and a hole in the spring, the 
advocates of this method either confusing or misunder- 
standing the difibrent actions of a spring in a barrel, 
where any adjustment of its strength is sought for in the 
mode of its attachment to the barrel, and one actuating 
a train through a fusee. 

A mainspring to act with a fusee is tapered from the 
outer end (the adjustment of the fusee enabling this to 
be done), so that when the outer end of the spring is 
fixed rigidly to the barrel, the coils of the spring when 
being wound round the arbor will fall away from the 
outer circumference separately, and wind and unwind on 
and off the arbor without the friction produced by the 


coils rubbing one another, as they do with a straight 
spring and a loose attachment at the bairel. 

Notwithstanding the growing disposition amongst 
modern English watchmakers to praise and adopt the 
going barrel, all the old French watchmakers and writers 
on the subject agi'eed as to the superiority of the fusee 
and the tapered mainspring with a rigid attachment 
to the barrel. 

M. Saunier, in his treatise, quotes various authors, 
and enumerates experiments on this subject made by 
eminent French watchmakers. He says that A. Breguet 
and U. Jurgensen made marine chronometers with large 
going barrels and long springs of convenient form, of 
which only a few coils, selected by test, were ever called 
into action. The spring used by Jurgensen in one of his 
marine chronometers without a fusee, which had a 
remarkably good rate, was very weak, and had a length 
of 3*5 metres (11 5 ftj. Its (the chronometer's) price 
was very high, and the maker did not care to make 
others at even that price. He says : "It is essential that 
great care be exercised in ascertaining the progressive 
force and in the construction of such a spring, in order 
that it may give a uniform pull and always keep the 
several coils apart, for otherwise it is impossible to avoid 
resistance caused by adhesion and clustering." He says 
again that it is difficult to apply, these long springs to 
pocket chronometers, because they cannot, from want of 
space, be of the requisite length and breadth ; and that 
M. H. Robert has obtained a sufficient approximation to 
uniformity in the force of a spring of ten or eleven turns 
in his going barrel chronometers, utilising 3*5 turns from 
the " point of winding up, when the last turn was out 
of contact with the barrel rim." Considering the great 
ease with which an ordinary marine chronometer main- 
spring with a fusee can be perfectly adjusted, without 
any chance of resistance from adhesion or clustering, it 
seems strange that these men, who were so well 
pcquainted with the superiority of the fusee, should have 


devoted so much time to experiments with going barrel 
chronometers, and it is a question whether this pei> 
sistence in trying to supersede the fusee adjustment in 
marine chronometers has not lost this branch of the trade 
to France. 

Now that the introduction of keyless watches has 
brought the going baiTel into common use, it would be 
foolish of English watchmakers to apply it only with 
a grudging acquiescence, or to make no effort to improve 
its construction. At the same time, it is well to remember 
that a great many experiments have been made by the 
French watchmakers referred to in a variety of ways, the 
outcome of which seems that a spring of moderate 
length set up about a turn, with something to spare, will 
have an easier action in the barrel, and be freer from 
adhesion and clustering than a spring of greater 
length, although the longer spring may give a more 
uniform pull if set up several turns. It is also worthy of 
notice that all the best Swiss watches have a rigid 
attachment of the mainspring to the barrel, Us, although 
the hook is in the barrel, and the usual oblong hole in the 
spring, the attachment is made rigid by the pivoted 
brace or post, which contributes greatly to the free 
action of the spring, and prevents to some extent the 
friction and adhesion from the coils rubbing against one 
another. The Swiss, by using a differently-arranged 
movement, obtain a wider mainspring than could be got 
in an English movement of the same thickness, and ~as 
there is no room to spare in pocket watches, it is of great 
importance to make use of all the barrel space availabla 

48. Number of Turns in the Spring. — To prescribe a 
certain number of turns of spring for a barrel would be 
quite misleading, as the number of turns of the spring in 
the barrel is no guide to the number of turns the barrel 
will make by the winding or unwinding of the spring on 
or off the arbor. 

M. Saunier quotes the following theorems from a 
work on mainspiings by Messrs. Koz4 (father and son), 


which was published in Volume II. of the " Revue 
Chronometrique " : — 

(1) A mainspring in the act of uncoiling in its 
barrel always gives a number of turns equal to the 
difference between the number of coils in the up and 
down position. He says : — "Thus if 17 be the number 
of coils when the spring is run down, and 25 the number 
when against the arbor, the number of turns in the 
uncoiling will be 8, or the difference between 17 and 25." 

(2) With a given barrel spring and arbor, in order 
that the number of turns may be a maximum, it is 
necessary that the length of the spring be such that the 
occupied part of the barrel (exclusive of that filled by 
the arbor) shall be equal to the unoccupied part ; in other 
words, the surface covered by the spring when up or 
down must be equal to the uncovered surface of the 
barrel bottom. 

A good deal of stress is laid by various writers 
on the necessity of a proper-sized barrel arbor ; 
but if the arbor used is too small, as it often is 
in fusee watches, when too thick a spring is used, the 
mainspring will break at the eye, unless it is made very 
soft at that part, when the only effect will be that it will 
bend round the arbor, acting as a larger arbor and 
reducing the acting length of the spring. 

The size of arbor found to answer best, allowing of 
the necessary length of spring and preventing too small 
a circle at the eye, is one-third the inside diameter of the 
barrel ; the arbor should be snailed, so that when the 
spring is wound on to it, it will take a spiral form, and 
not be distorted, as it would be by winding it on a 
circular arbor. 

In order to diminish the friction of the coils in a 
going barrel, mainsprings have been made recently with 
the outer coil curved backwards, so that the spring when 
unconstrained takes a form something like the letter s. 
This spring is made with a view to the better separation 
of the coils upon the spring's unwinding, as the outer 


coils will fall more readily away from the inner ones 
towjirds the. (mIj^o of the barrel when the spring is bent 
in til is way than when it is straight or of the usual form. 
It is said to be freer in the barrel, but liable to break. 

49. Spring Attachments. — There have been so 
many inventions for equalising the pull of the spring 
by the mode of attachment to the barrel, that a 
few of thein may be noted, although, as abready 
Ktat(Kl, they have been all abandoned for the rigid 
attachment in the best work by the Swiss, who 
have had the most interest hitherto in adopting any 
improvement applicable to the going barrel. The 
Americjins are inventing over again and patenting some 
of the obsolete French inventions. A mainspring attach- 
ment attributed by M. Saunier to M. Philippe, but 
claimed by a dozen others, is obtained by bending the 
end of the spring backwards in the form of a hook, and 
forming four nicks in the inner edge of the barrel, in 
which this hook will catch and hold, unless unusual strain 
is put upon the w^inding arbor, when the hook would 
slip from the nick and catch in the next one. It is 
claimed for this plan that no stop work is required, but 
it is really quite useless. In the first place, the rim of 
the barrel would require to be two or three times 
the usual thickness to admit of the nicks, or grooves, 
being cut deep enough, while the bent-over end of the 
spring standing out from the side of the barrel would 
take up three or four turns of spring room. Secondly, 
after the spring hook had been drawn across the nicks a 
few times (if it did not break before), it would not hold 
in any one of them. And yet this attachment is quoted 
as one of the improvements in modern watches. 

Another mode of attachment, accredited by Saunier 
likewise to M. Philippe, but having many other 
claimants, is that of having an elastic ring sprung into 
the barrel, the end of the spring proper being riveted 
to it at one end. This ring fits the barrel sufliciently 
tightly to resist being moved during the winding of the 


spring, but when the spring is fully wound, and the 
outer coil being pulled towards the centre of the barrel, 
it will give sufficiently to prevent the spring from being 
broken. This method offers attractions in the shape of 
rigid attachment, getting rid of stop work, &c., but is 
open to the same objection of taking up too much space 
as the previous one, the spring band being necessarily 
considerably thicker than the mainspring. An improve- 
ment on this invention is that of substituting a piece of 
mainspring for the spring ring ; this piece is riveted on 
to the end of the mainspring outside it, and running the 
reverse way, or, as it were, doubled back upon it ; at the 
other end it has a hole which is held by the usual pin in 
the barrel ; when the spring is wound up the piece falls 
towards the centre, but offers sufficient resistance to the 
winding to be a substitute for a stop. 

But practically there are only two modes of attach- 
ing the spring to the barrel that have obtained general 
approval, namely, the rigid attachment formed by the hook 
in the mainspring and square hole in the barrel, and the 
loose or flexible attachment of the hook in the barrel 
and the hole in the end of the spring. The Swiss in all 
their best watches have converted the latter attachment 
into a rigid one by the use of the pivoted brace. If 
there is any reason, apart from it being their ordinary 
custom, for adopting this way of getting a rigid fixture 
of the end of the spring, it is that the mainspring could 
be more easily changed if too weak or too strong, or if 
the spring were to break, than if the hook were 
riveted into it, as in the latter case a new hook would 
require to be made ; but the hook in the spring takes 
up less room, and this mode of attachment is un- 
doubtedly the best. 

60. Q-oing-barrel and Stop Work. — Fig. 6, a, is a dia- 
gram of the ordinary going-barrel and " Geneva " stop 
work. Its application to all the watches made out of Eng- 
land, and to nearly all the keyless watches made here, 
renders it familiar to every watchmaker, and any elaborate 



[Ou*. TL 

description of it is uuneoeeatay ; and, as Ute principle td 
the action of the mainspring in the barrel has already been 
discuBsed, the only feature that need be noticed is the stop 
work. The stop work shown in the figure is applied to 

Tig. 6— 

nearly all going barrel watobes ; but this and every form 
of stop work so applied has the great defect of reducing 
the height of the barrel, necessitating a narrower, 
thicker, and shorter spring, and thus increasing its 
liability to break. There is also the further defect of 
the stop drawing away the oil from the hole and pivot to 
which the stop finger is fitted, as, to save nwm, the 
finger must be fitted close to the barrel. The Swiss 


always fix this stop work on the barrel cover, but some 
of the Lancashire movements have it on the barrel This 
arrangement is bad, for these reasons : — The barrel hole 
being in the same plane as its teeth, or wheel, there is 
more friction in that hole than in the hole in the cover, 
and, therefore, the pivot should be as long as possible and 
have oil kept to it ; the sinks turned out to receive the 
stop work necessarily weaken the part of the barrel to 
which it is applied, and render it more liable to be spoilt 
than the un weakened part, and any injury to the cover 
is easier remedied than if it occurred to the barrel (In 
the event of a barrel being damaged from this cause, it 
can be rectified by turning out the bottom, snapping a 
new one into the groove, similar to the one into which 
the cover is snapped, and riveting over). 

With a view principally to remedy these defects of 
the ordinary stop work, and also to obtain an approxi- 
mate adjustment of the mainspring, I devised the barrel 
represented in Fig. 6, b. 

Two slots are cut tangontially, as shown in the figure, 
in the barrel and cover, into which the pivots on the 
piece c project. Tlie piece c is riveted to the end of the 
spring with one rivet only, this allowing of its bearing 
equally on either slot in case they are not exactly 
parallel to one another, and half the thickness of the 
pivots is filed away to prevent loss of spring room. 
When the spring is wound round the arbor the pivots 
traverse the slots inwards, and, when they reach the 
bottom, make a strong and most eff*ective stop. 

The number of turns the spring is required to make 
is regulated by the length of the slots ; if they are too 
long the spring will be wound up on itself, and, although 
the stop will be perfect, the friction of the coils of the 
spring on each other will be greatly increased (an evil 
that is not always considered in the going barrel, as in 
many cases the whole length of the spring is used). 

This bari^l has the advantage of admitting a wider 
mainspring and longer pivots ; there is nothing to draw 


away the oil from the holes, and there is besides a perfect 
adjustment of the mainspring for nearly four turns. 

5 1 . Mainspring Making. — From somecause noteasilj 
explained, watch mainspring making has almost died out 
in England, and I should be afraid to say how many 
of the Hj)rings applied to English watches are made in 
Switzerland. No country has greater facilities for draw- 
ing and rolling steel than this, and most of the steel 
used in Switzerland is English, yet we have allowed a lucra- 
tive industry to slip away from us. We have, however, 
still the monopoly for marine chronometers, and of the 
best hammered and tapered springs for fusee watches, 
springs hammered or forged from wire being more elastic 
than those cut from rolled sheet steel. 

The process of making a chronometer spring is so 
nearly identical with that of making a watch spring that a 
description of the former will suffice for the two. 

These large springs are cut from rolled sheet steel, 
which must be of the best quality, with a circular 
cutter ; the bench on which they are cut has an adjust- 
able bar, so that the springs may be cut to any required 
width. The roughness left from the cutter being taken off 
the edges with a file, the strip of steel is then bound 
round with binding wire, and wound up into a spiral 
six or eight inches in diameter (the binding wire keeps 
the coils separate and the steel from scaling). It is 
then heated in a close furnace, and^ when at a proper 
heat, dipped into melted tallow or oil ; as steel will not 
all harden at the same temperature, a good deal of the 
art of spring-making lies in the careful hardening. 

When the spring is removed from the tallow and 
the wire taken otF, it is hard and very much distorted; it 
is then drawn over an iron case, heated with gas until 
it becomes of a deep straw colour, and in this first pro- 
cess of tempering it is brought nearly straight, and soft 
enough for further manipulation. It is now fixed in a 
frame, somewhat resembling the frame of a large bow 
saw ; this frame moves along a bench on rollers, and 


has a screw at one end for drawing the spring tight as 
the heat to which it is subjected lengthens it The 
spring is again placed over the hot iron case, and 
when the part at the end that is first submitted to the 
heat becomes of a light blue colour, it is drawn gradu- 
ally over it until it is all of the same colour and temper ; 
the rollers on the frame enable the operator to move it 
quickly if the colour is coming too rapidly. The spring 
is kept quite tight in the frame during the process of 
tempering, and when it cools all the kinks have dis- 
appeared, and it is quite flat and straight. It is next 
fixed up in a frame similar to the last, having one end 
fixed and the other adjustable, and worked smooth with 
lead clams and coarse emery. After all the marks have been 
removed, it is tapered by working the clams backwards 
and forwards, beginning close to one end of the spring, 
and increasing the length worked upon by an inch each 
time the clams are drawn to and fro, the operator 
gauging the thickness as he proceeds, and finishing with 
wooden clams and fine emery. 

Notwithstanding the constant gauging of the spring 
during the process of tapering, the amount of tapering 
and its accuracy are tested at the last by bending the 
spring into double loops from end to end, and seeing 
that these loops grow gradually larger or smaller, 
according to the end first tested : this testing requires 
skill and judgment. When the spring is finished Vith 
the fine emery, the eye is made at the thin end, and a 
small part of its length softened for bending round the 
barrel arbor (if it were left as hard as the rest of the 
spring, it would break by being bent into so small a 
circle) ; it is then wound up two or three times, and if it 
forms a perfect spiral, it is passed, but if any of the 
coils are closer together than the rest, it is rejected, as 
this shows it is not of uniform temper throughout. 

Straight springs are so much easier made than 
tapered ones that they can be made much cheaper, and 
it can only be want of enterprise on our ^^y^ ^\sa^ 


prevents us from supplying other countries with main- 
springs as we have formerly done, instead of importing 
the greater part of what we use. 



52. Chronometer Finishing. — The principles which 
^vem the action of wheels and pinions are all that 
an intelligent workman needs to understand — apart 
from the practical details — in order to be able to 
finish well ; but the proper application of these practical 
details can only be indicated in a book, and it requires 
long experience at the bench before a man becomes a 
good finisher. 

Chronometer finishing is only good clockmaking, but 
the work is so much finer that clockmakers rarely take 
to that or any other branch of chronometer making. 

The barrel and fusee of the chronometer being large 
work, what remains for the finisher to do to them is 
usually done in a small foot-lathe, or in a hand-lathe, 
or throw. The pinions are sometimes pivoted and finished 
in a throw, but for the most part this work is done in 
the turns, there being no machine that wiU ennoble 
one to turn an object as true as it can be tv/rned on iti 
oum centres, rotated with a bow in the common turns, 
especially if it be a thin-tempered arbor. 

There is a little to do to the barrel when it oomes 
from the movement maker, namely, to smooth the pivots, 
and free the barrel on the arbor and let it into the frame, 
and hook in the mainspring. 

63. Hooking in the Spring. — The springs have always 
been until recently fixed to the barrel, with a square 
hook riveted in the spring and fitting a corresponding 
hole in the barrel ; but lately the Swiss mode of a hook 


in the barrel and a hole in the spring has been adopted, 
because it is easier done, and if a new mainspring is re- 
quired there is no new hook to be made. 

Considering the necessity for perfect adjustment of the 
mainspring of a ship's chronometer, and the trouble taken 
by a good mainspring maker to taper the spring in order 
that the inner coils, when being wound round the barrel 
arbor, may fall away from the outer ones without friction, 
the advantages of a firm attachment of its outer end 
to the barrel will be at once recognised. This hook is 
better than the barrel hook usually found in watches, 
there being more room for it, and it being turned of the 
form of the head of a screw for wood or of a button ; 
but whatever its form, it is not so good as the hook in 
the spring, and it takes up a portion of the spring room, 
since it must project beyond the thickness of the spring 
which it holds. (For making a barrel hook, see page 

54. Planting the Wheels. — After freeing the great 
and maintainiiig-power wheels with regard to one another, 
and smoothing the holes in them and the arbor, upon 
which they will be found very tight, the spiral should be 
cut for receiving the chain, the number of turns to be 
cut on it to make the chronometer go the required number 
of hours being ascertained, as directed at page 60. The 
pin holes through the arbor and the collet that keeps the 
maintaining-power and great wheels in their places should 
be broached together, and the arbor and holes smoothed 
from burrs. The wheels may be freed by either turning 
a little off the face of the collet or out of the centre 
of the great wheel, until they move easily with a pin 
fitted in. The wheels should move easily, and yet have 
no side play or shake ; as if the collet is pinned on so 
that they are stiff to move, the chronometer is hard to 
wind, and the maintaining-power spring is prevented 
from acting. The centre wheel is pivoted first : it must 
be planted in the centre of the frame, and just be free of 
the pillar plate. The fusee should be planted so that 



the great wheel makes rather a deep depth with the 
centre pinion, and is free of the barrel. 

55. Stop and Spring. — Tlie stop which catches the 
fusee hook when the chain is all wound on should next 
be tiled up and adjusted, the pin-hole through the stop 
and stud being broached with the stop held close to the 
plate at the point. When a pin is fitted, the back of the 
stop can be filed away until it just rises sufficiently from 
the plate to allow the hook to pass between its point 
and the plate ; it must then be filed to length. When the 
stop forms a right angle with the face of the hook, it 
will be the right length, the face being radial to the 
fusee centre ; the hole in the blade of the stop must be 
broached a little larger than the pin, in order to allow the 
stop to rise from the plate. 

The stop spring must not be strong, and should only 
touch the plate where it is screwed to it, and a notch is filed 
in the stop near the blade to give the spring freedom. A 
part of the stop where the chain presses has to be 
thinned so that it may not be pressed in too soon. This 
part may be roughly got by twisting the cord of a small 
bow once round the smaller end of the fusee, putting the 
fusee top pivot in its place, and holding the bow so that 
the cord is parallel to the plate, and in the direction of 
the edge of the barrel from which the chain would be 
unwound. The amount of thinning it may require can 
be done without putting the barrel and fusee into the 
frame. When the mainspring has been adjusted, the action 
of the stop can be tested, and any correction may be 
made before it is finished. 

56. Adjusting - rod. — The adjusting-rod shown in 
Fig. 7 is somewhat different from the ordinary one used 
by finishers for testing the pull of mainsprings, &c. In 
addition to the sliding weight, the jaws which grip the 
squares can also be shifted, being fixed at any required 
distance along the rod by means of a thumb-screw at the 
back. This enables it to be used for almost any spring, 
from the weakest to the strongest 


After adjusting the mainspring, the adjusting-rod 
should be put on the fusee square before the weight is 
shifted, and the great wheel held in the hand in such a 
position that the weight will draw the pin in the 
maintaining spring to the other end of the slot in the 
wheel through which it projects. If the weight of the 
rod does not draw the pin forward, the maintaining- 

Fig. 7.— Adjusting-rod. 

power spring is too strong, and must be weakened ; the 
mainspring must be strong enough to draw the main- 
taining power spring as far as the slot will let it go when 
the power of the spring is pulling on the fusee. A brass 
cap is let on to the fusee square, to keep dirt from getting 
into the hole when the piece is wound, and a pipe is 
generally screwed on to the plate round the winding 
square, and let through the bottom of the brass box, to 
keep dirt and damp from the chronometer. 

57. Finishing the Wheels and Pinions. — The pillar 
plate has a circle cut out, and a bar screwed over it on 
the dial side to receive the bottom and seconds pivots of 
the third and fourth wheels ; this arrangement enables 
the third wheel to be run under the centre wheel, and 
prevents the seconds pivot from being too long. The 
fourth pinion must be pivoted close to the bar, and must 
have a deep hollow cut in its face, to prevent the oil 
from being drawn from the pivot ; the third wheel must 
be kept free of the centre wheel, and the pinion must 
have a hollow cut in its back or rivet, as must also the 
centre pinion. 

All the wheels have to be polished. If they are true 
on the sides, they are first stoned with a smooth Water- 
of- Ayr stone on a cork until they are quite flat, and then 
placed in the turns and polished circularly with an ivory 
polisher, and a paste made by rubbing two pieces of 


blue-stone together with a very little oil, and finished 
with a slip of boxwood and a few rubs with a piece of 
willow at the last ; if the willow is used for too long a 
time, the arms of the wheels will be rounded. 

Although polishing the wheels is the usual way of 
finishing them, stoning them very flat and smooth and 
electro-gilding them will give them a much nicer appear- 
ance, and keep the brass from tarnishing. 

The pinions and arbors are highly polished; some 
finishers burnish the arbors, but a high polish can be got 
very quickly with a zinc polisher and diamantine. 

The faces of the third and fourth pinions are finished 
with the ordinary facing tool, but as the large pivot on 
the centre arbor precludes the use of such a tool, it is 
faced square down to the arbor ; the pinion is placed in 
the turns, and small turns that fit into the rest holder 
carry a roller mounted on an arbor ; this roller is brought 
to bear against the face of the pinion, and the pmion is 
rotated backwards and forwards with a bow. The roller 
first used is steel, to bring up the face flat and square, 
after which soft metal rollers are used for finishing. 

68. Jewelling. — The third and fourth holes are 
usually jewelled, although some makers object to jewel 
these holes, as the pivots are apt to become black ; but 
if jewelling is indispensable in the escapement holes, it is 
also necessary here, and if the holes are good and well 
made the pivots will not wear. 

The fourth wheel hole in the upper plate should be 
jewelled with an endstone, although it is seldom done, 
as a pivot of equal size is stronger if conical than one 
with a square shoulder, and the pivot will run with less 
friction on the end than on its shoulder. The oil also 
keeps better in a covered hole. 

69. Spotting the Plates. — The spottmg of the plates 
is a branch in itself ; it is done in an engine resembling 
a wheel-cutting engine ; after the plates are polished with 
rotten-stone, the plate or piece that is to be spotted being 
fixed to the dividing plate, a small hollow ivory point 


charged with oil-stone dust or emery and oil is attached 
to a jambed arm. This point is brought into contact with 
the plate while it is rotating, the plate being shifted 
after each spot, and circular or geometrical patterns 
marked on it as arranged on the dividing plate. The steel 
work and the screws are blued, but it is not thought safe to 
harden the large screws, since, if a screw head broke off 
and stopped the chronometer, the result might be serious. 

60. Watch-finishing. — ^The various operations that 
have hitherto been combined under the name of watch- 
finishing I have to some extent treated of separately, for 
I think that, if ever watchmaking resume its former pro- 
portions in England, some more economical and rational 
system must be pursued in manufacturing than that 
now in vogue. No better way could be devised of 
making a young man a thoroughly practical watch- 
maker than by apprenticing him to a finisher first, 
and afterwards to an escapement-maker. But the time 
occupied in waiting for work that must be done by 
othera — such as jewelling, engraving, gilding, &c.— 
the diversity of the work — from pivoting and planting the 
barrel and fusee, to making and planting the stud and 
index, &c. — and the tools and appliances necessary to 
enable one man to do these things in the best manner, 
are so considerable that it has not been possible for even 
the most competent workman to obtain sufficient 
remuneration for the time and labour expended on the 
best work; while men with less ability and principle 
have had a constant temptation to shirk such parts of 
the work as were most troublesome to them, or in which 
bad workmanship was more difficult to detect. The 
practice hitherto has been for the finisher to take the 
movement after the dial and escapement were made, and 
to return it to the manufacturer gilt, proceeding in the 
following manner : — 

61. Testing the Wheels. — The first work is to get 
the wheels ready for gilding, and it frequently happens, 
in movements that are not of the best quality, that they 


require to be tliiiined ; in that case thej should be placed 
in the frame and the space for them noted. I have often 
found finishers thinning wheels from habit when there 
was no occasion for their doing so. 

The pinion arbors must be shortened and the pinions 
and wheels got true; the wheels must be made perfectly true 
on the sides, square and flat (first on a cork with a piece of 
Water-of-Ayr stone and afterwards rubbed circularly in 
the turns) ; and hollows must be cut in the wheels to per- 
mit of the rivets of the pinions being polished without 
touching them when they are gilt. It is sometimes better 
to make the lower pivot on the centre wheel arbor before 
the wheels are gilt, as then the barrel and fusee can be 
worked into their places. 

62. Running in the Wheels. — The centre pinion 
has an arbor fitted into it projecting at each end, 
far enough to put a ferrule on it ; a deep hollow is 
cut in the back of the pinion, and the pivot made as 
close to the wheel as possible. The hole in the plate 
should be well chamfered and a turned stopping of 
good brass put into it, the stopping being allowed 
to project on the dial side of the plate ; the centre 
wheel should always be kept a little below the plane of 
the plate, to give it sufficient freedom from the barrel 
and great wheel ; the fusee and balance being put into the 
frame, it will be seen if there is sufficient freedom 
between them. If the balance is a plain one, and is quite 
free of the fusee brass, the chain will be free also, 
but a compensation balance should have more room 
allowed it, as the rim may be bent after it is cut by 
careless handling; if the movement is right in this respect, 
the fusee can be pivoted and planted. The barrel hole in 
the pillar plate need not be stopped if the barrel is a 
proper height in the frame, and a sufficient distance 
from the third wheel arbor and the great wheel. The 
upper fusee pivot should be as long as possible. Move- 
ment makers do not leave it long, as it is a common 
practice to cut and polish a hollow in the fusee arbor ; 


but there is not space for this 'with a long pivot, as there 
is a hollow cut in the reverse side of this shoulder in the 
ratchet wheel to admit the short projecting pi])e of the 
great wheel. Since, then, thei*c is not thickness enough for 
a hollow on both sides and space for a lon^ j)ivot, it is 
much better to leave the shoulder s^piare and turn the 
fusee piece sloping from the hole to make the bearing 
light, or take a large comer off the shoulder, jjolishiiig 
the slant, giving the required bearing, as with other 
square pivots. It is customary to |)olish the inner and 
outer edges of the fusee cap before snailing it. The inside 
edge, after being turned to the required angle, is polished 
with a tool, similar in principle to an ordinary pinion 
facing tool, but turned or filed to fit the angle or edge 
of the cap ; the outer edge is more diiHcult to do ; it is 
always polished, even in common watches, but there is 
no reason why it should be, as it is not always orna- 
mental. The outer edge being sloped off in the turns till 
its edge comes to the plane of the top part of the brass 
cone, is polished with a broad polisher, one side of which 
resting on the brass, enables the edge of the cap to 
be polished quite flat ; the back of the hook is Hied to a 
corresponding edge and polished on a cork. Some 
finishers take out the arbor and polish it all upon a 

The pin-hole in the collet and ar})or should be 
broached until it is soimd with the fusee together, and, 
if the wheel and collet are too thin to bear reducing to 
give the necessary freedom, a piece of tissue paper 
placed between them while the hole is broached will 
ensure it. This freedom of the great wheel is a point 
generally neglected, and is of much more importance 
than the polishing of the edges of the cap. 

63. The Top Plate.— The depth of the great wheel 
and centre pinion being determined and marked, and the 
stopping for the fusee hole put in, the t )p plate should 
be turned out for the fusee piece in the following 
manner: — 


Place the pillar plate in the mandrel with the centre 
in the fusee hole, and peg the hole (pegging the hole is 
the only, method by which a perfect upright can be 
obtained in the mandrel, and is the usual way of getting 
it). Take a long peg and cut a fine point to it, and, 
withdrawing the mandrel centre, place the point of tlie 
peg loosely in the hole. Then by bringing the turning 
rest up to about an inch from the plate, level with the 
hole, resting the body of the peg on it, and gently 
turning' the mandrel, it can be seen whether the hole is in 
the centre exactly ; if it is not, the end of the peg pro- 
jecting over the rest will move up and down, when the 
dogs must be slightly loosened, and the plate tapped 
gently until the hole is in the centre. 

When the hole is right, tighten the dogs, screw on 
the top plate, and turn out the place for the fusee piece. 
An exactly similar course must be followed in putting in 
the holes in the top plate for the centre wheel and 
barrel ; but the upright of the centre wheel is of so mucli 
importance that a hollow stopping should be put in, and 
the inside of the hole turned out with a fine cutter in 
the slide rest of the mandrel, to nearly the size of the 
pivot. It may not be requisite with a good mandrel to 
peg the hole before turning out the place for the fusee 
piece, but it is quite necessary to do so when turning 
out the pivot hole in the piece itself. 

64. The Barrel Arbor. — If the pivots of the barrel 
Jirbor are the proper shape (which they generally 
are now in the best movements, and certainly ought 
to be), the pivots and holes will only require smooth- 
ing, and the barrel freeing on the arbor. Instead 
of adopting the usual course of turning away the bosses 
in the barrel and cover to reduce the rubbing surfaces, 
a deep hollow should be turned, and a shoulder formed 
on each side of the arbor of a sufficient width, and the 
bosses should be left on the brass as large as possible. 
It has not been the practice to snail the barrel arbors 
of fusee watches, as there was no trouble with the adjust- 


ment of the mainspring, English springs hoing taperod 
Mid generally filed thin at the eye, but the arbors 
should be snailed (and they probaV>ly will be now by the 
Tnovement maker), and the hook should not project 
beyond the thickness of the spring. 

65. Hooking in the Spring. — A spring of the pro- 
per length and strength being fitted to the barrel, it 
should be hooked in as follows : — With the 8i)ring in the 
barrel, drill a small hole in the barrel a little nearer to 
the bottom than to the cover, so that it may be in the 
centre of the inside of the rim, and within half-an-inch 
of the end of the spring ; the diill will mark the spiing. 
Remove the spring from the barrel, and broach the hole 
to nearly the size it is intended to leave it at an angle 
of 46® ; file this hole with a small square file, having a safe 
edge, mitil it is oblong, and, when it is of the required size, 
finish it with a drift. If the hole is in the middle and 
has not been drawn to either side of the barrel by filing 
it, the end of the spring can be softened, as far only as the 
hole for the hook, until it is a light blue colour, but it 
should not be made softer; the hole is drilled by 
gripping the spring against a piece of wood or brass in 
a small hand-vice. 

The hook should never exceed one-third the width 
of the spring, a larger hook unnecessarily weakening the 
barrel and holding no better. A piece of rectangular 
steel is fitted to the hole in the barrel, and a mark is 
scored on it with a sharp point along the inside of the 
barrel. This mark will give the angle at which the 
shoulder should be made ; if the angle is less than 45^, 
the hook will probably draw out if the spring is fully 
wound round the arbor. The pivot on the hook is often 
made by filing, but it is much better and quicker to 
make it with a cutter. As the strain comes on the back 
of the hook, the pivot should be kept as near the front as 
possible. The cutter is a piece of steel with a hole 
drilled in it and fine teeth cut on its face ; it may either 
be a piece of wire with a ferrule on it, or it may be fitted 


to a drill stock. In the latter case the hole should be 
broached from the back to enable the cutter to free 
itself, but as the pivot need not be long, this is not 
imperative ; the cutter should be slightly convex on the 
face to ensure a firm seat for the hook when riveted. A 
point or centre being left on the steel which is to form 
the hook, a few strokes of the bow will form the pivot, 
and the angle of the hook when the pivot is made should 
correspond with that of the hole in the barrel. The pivot 
is fitted by broaching and chamfering the hole in the 
spring, and riveted by gripping the steel in a blunt pair 
of nippers and screwing them up in a vice ; the rivet 
should be left long, and made with a few strokes of a 
rather heavy hammer. There is a good deal of strain on 
this hook, and for that reason the attempt (generally a 
failure) at using a hook a second time should never be 

About three-eighths of an inch of the spring should be 
left beyond the hook, and this end must be filed away to 
a knife edge, the thinning to commence at the hook. If 
the spring is left the full thickness at the end and an 
unusual strain is put upon it, it will break across the 
hole where it is weakest, or the hook will draw out. 
Most watch jobbers know this, and generally make the 
end of the spring so soft that it bends, and all the 
advantage of a rigid attachment is lost. 

If the spring is fitted to a new barrel, the hook can 
be filed down from the outside of the barrel, and finished 
on an arbor in the turns ; if, however, the barrel is gilt 
or polished, the hook must be finished before it is finally 
put into the barrel ; it is brought to height by trying it 
in the hole from the outside, and filed down and finished 
on a cork ; the hook must not project beyond the barrel. 
If there is any difficulty found in getting the height of 
the hook by trying it from the outside, the thickness of 
the rim and mainspring can be measm^ed separately by 
a douzieme gauge; the thickness of the two together 
will be the thickness of the mainspring and hook. This 


is the best way of hooking in tlit; ]iuiinH])nii;^ of a 
fusee watch, but there are various reasons fur depart- 
ing from it in going barrel watclies, one of the principal 
being that a flexible attachment of tlic; s])ring to tlic 
barrel to some extent affords an adjustment ; Uu; pull ot 
the inner coils is not affected by the attachniont, but, 
when the outer coil is called into action, it falls away 
from the barrel with less resistance when the attiiclinient 
is flexible than when it is rigid. To satisfy myself of 
this, I made the following exi)erinient, wliich I think 
conclusively proves my assertion to bo correct. Having 
carefully adjusted the mainspring of a " two-day " marine 
chronometer which had a hook in the barrel, I removed 
the hook in the barrel and substituted the s<{uare hook 
in the spring. Upon again trying the adjustment, 
I found that to obtain an adjustment the spring had to 
be set up considerably more than it was bcifore, and the 
weight of the adjusting rod altered, as the pull of the 
spring was greater throughout. 

Again, there is greater uncertainty in fitting a spring 
of the proper strength to a going barrel than there is to 
a barrel with the fusee, in consequence of the inequality 
of the arcs of vibration of the balance during the un- 
winding of the spring of the former ; for if the balance 
vibrates over a turn when the watch is nearly down, it 
will be liable to bank with the least external motion 
when fully wound, especially if the lever escapement bo 
a good one ; and it is much less convenient to change a 
spring with a hook in it, than one that has only a square 
hole punched out of the end. 

As the springs are now nearly always fixed in going 
barrels with a steel hook in the barrel, some system 
should be observed in making them. In English 
watches the greater part of the mainspring is brought 
into action, and, as no pivoted brace is used, the hooks 
must be made more secure, otherwise the spring will 
either slip off or pull the hook with it. 

If a small screw plate is made with two or three 


holes of the sizes required, drilled and tapped at an 
angle of 45°, the pin or hook can be tapped, and the end 
of the screw filed up into a hook of the required 
form, and to the proper height ; or it can be 
tapped in an ordinary screw plate, and afterwards 
screwed into the slanting hole and finished, but it is 
necessary first to screw the pin into the barrel, and raark 
for the inside of the hook close to the rim, so that, when 
the hook is in the right position, the screw part of it 
will be flush with the inside of the barrel and quite 

If this hook is properly formed and at a sufficient 
angle, the spring will not draw off, even if it be left very 
little higher than the thickness of the spring ; the end of 
the spring in this case is made softer than if it had a 
rigid attachment. The hole should be oblong and 
perfectly square at the outer end, bevelled off from the 
inside to take a good hold of the hook, and there must 
not be any length of spring left beyond the hole but 
what is required for strength, as, if there is, it acts as a 
lever and pushes the spring off the hook, even if thinned. 
On the other hand, if a round hole be made in the spring, 
the strain draws the middle of the outer circumference 
of the hole to that of the hook, and unless the two 
correspond or are exactly in the middle of the rim, the 
spring will press either the cover or the bottom of the 
barrel, and add considerably to the jerking action of the 
spring, so familiar to watchmakers in the winding and 

Another, and better, although more difficult way of 
hooking in the spring is to screw the hook in from the 
inside of the barrel. . The chief trouble attached to this 
method is in getting the hook square to the barrel rim 
when screwed home to the shoulder ; it must be tried first 
and marked for the square, a little being allowed for it 
at the last to be screwed home veiy tight, as in marking 
for dog screws. In addition to the other reasons for 
not hooking in the spring as in fusee watches, the 


extra width of the great wheel teetli attached to the 
barrel has to be considered, which would necessitate in most 
cases this hook being placed at one side of the spring 
instead of in the centre of it. 

66. Adjusting the Chain. — When the spring has 
been hooked in, the fusee piece made and planted, and 
the fusee and barrel run in, the chain should be 
attached and the stop work filed up and adjusted. 
Care must be taken that the hole in the barrel for the 
chain-hook be placed so that the chain will be just free 
of the plane of the top plate ; the hole may be drilled 
anywhere in the barrel, provided a little is filed out of 
the cover to free the end of the hook should it project 
into the sink, which it generally does. If the chain is 
too far from the plate, the second turn will probably ride 
on the top of the first, hence the necessity of keeping 
it as close to the plate as possible, and to avoid this 
riding of the chain it must not be left too long. The 
barrel hook should come to the outer edge of the top 
plate when the chain is wound up. 

The hollow in the stop may be filed out to nearly its 
proper thinness (see chronometer finishing, page 82), 
but in no case should this process be considered sufficient. 

If watch finishers, or even examiners, had thoroughly 
understood the uses of the adjusting rod and freely used 
it, half the faults that have been found with the fusee 
watch would have been avoided. It is not , that the 
actual adjustment of the mainspring is of such import- 
ance at this stage, but, with the fusee and barrel only 
in the frame, any fault that may afterwards give trouble 
can be seen and removed. If the chain does not lie 
square on its edge on the barrel, it should not be used, 
as no filing will make it do so ; it ought to lead freely 
into the grooves of the fusee, and not touch the stop the 
turn before the final one, and, when it does touch it, it 
should bring it close to the top plate as the hook of the 
fusee cap comes to it ; at all other times the stop should 
be quite free of the hook, and not rise high enough for 


the end of it to come in contact with the chain on the 

If the sinks for the barrel and fusee in the top plate 
are turned deep, it is sometimes difficult to prevent the 
chain from riding on the barrel hook ; to avoid this, 
accordingly these sinks must not be turned too deep. Too 
much endshake to the barrel or fusee will also cause 
this to occur, and the stop to act improperly, therefore 
there should be only just freedom before gilding. 

The detent should be pivoted without endshake, 
ratcheting in the steel wheel and just free of the great 
wheel teeth. 

67. Pivoting. — When the wheels are gilt, the 
pinion leaves are polished out with a piece of hard 
wood and red stuflf, as the gilding discolours them. 
The centre wheel, having already the hole in the 
pillar plate, should have the top pivot made below 
the top plate and a stopping put into the plate to 
meet the shoulder; by so doing, a longer pivot is got, 
and the square for setting the hands can be sunk 
in the plate. This is more necessary in a keyless watch, 
the case coming so close to the plate that scarcely any 
square can be left ; in some cases, no square is left, and 
the piece that carries the cannon pinion has to be pushed 
out to remove the pinion. If the centre pinion is too 
long, it occasionally comes in contact with the steel 
wheel of the fusee ; in that case the pinion should be 
shortened before the top pivot is made. 

As the pivot holes in the bar for the third and fourth 
wheels are generally jewelled, the bar need not be very 
thick, and should be reduced from the inside to get as 
much room for the wheels as possible. The fourth wheel 
pinion must be shortened until the wheel rests on the 
pillar plate, and the pinions, having previously been 
polished, may be pivoted; the fourth wheel should have 
the freedom divided between the plate and the escape 
wheel, a deep hollow being first cut in the face of the 
pinion, as the arbor beyond the pinion is so short, that 


without this hollow the pinion would draw the oil from 
the pivot hole. If there is much freedom in the sink for 
the third wheel, the back shoulder of the arlx)r may 
be left longer, as it is only necessary to keep the third 
wheel free of the centre wheel. The pivots of the third 
wheel should be of equal thickness ; the seconds pivot 
of the fourth should be thicker than the top pivot, and 
if it is made straight, should have the end tapered after- 
wards, as it is not possible to fit a hand on it otherwise 
that will not get loose in a very short time. Making 
nice pivots has always been what the workman was 
most proud of, and, as it is only after much practice 
that a large square shoulder can be made, these shoulders 
have often been left too large because it was thought 
they looked handsome, and wlien pivots were often made 
small to prevent friction, the shoulders were left large 
enough to cause three times the loss of power that even 
the largest pivots would have done. If the holes are 
jewelled, the pivot shoulders can hardly be made too small ; 
for, although the jewel holes are sometimes sloped off to 
prevent the adhesion caused by the contact of the flat 
surfaces, the precaution is of little use if the oil becomes 
thick, as the oil adheres to the rounded face of the hole 
much in the same way as if it were flat. 

When the fourth pinion is pivoted and ready for 
running in, the depth with the escape pinion can be got 
in the depth ing tool, and marked on the dial side of the 
bar; this depth must be taken from the escape wheel 
jewel hole with the endstone off. When the score is 
made on the bar the dial should be put on and the 
seconds hole marked so that the seconds pivot may 
come in the centre of the hole, or as near to it as pos- 
sible. Some finishers put in the bottom holes only ; 
but it is not possible to be quite sure of the correctness 
of a depth until the wheels ai*e run into the frame ; 
therefore a hole should be put in the top plate for the 
fourth upper pivot, and the depths and the heights 
be verified. 


The third wheel depths are pitched last, that with 
the fourth pinion being ascertained and marked first, and 
afterwards that of the third pinion with the centre wlieel. 
68. Motion Work. — The branch of watchmaking called 
motion making has all but disappeared, and the greatest 
advocates of division of labour will scarcely regret it. It 
was the old custom to have the motion work — i.e., the 
cannon pinion, minute and hour wheels, and set-hand 
piece — partly made and the wheels planted, before the 
watch was finished or the centre wheel run in. These 
wheels, &c., were usually inferior to the wheels in the rest 
of the watch, and the principle was bad, as there were 
always more leaves in the pinion of the minute wheel 
than there were in the cannon pinion. The finisher is 
now, however, or ought to be, supplied with the motion 
wheels, and when the centre wheel is planted, a proper 
depth can be made with the cannon pinion and the 
minute wheel. The stopping for the centre wheel is left 
projecting and should now be turned in the mandrel into 
a pipe ; the pivot should project well through the hole, 
and the cannon pinion, having a greater number of 
leaves than the minute wheel pinion, will be large 
enough to have a square sink turned out of its face, to 
free the pipe left projecting. This is the finisher's work, 
and can only be well done at this stage, as leaving it for 
the examiner to do after the pinion is polished and the 
frame gilt, is doing the work twice over and doing it 

The minute wheel depth with the cannon pinion 
should be as deep as is consistent with perfect freedom, 
as should also that of the hour wheel and minute wheel 
pinion, in order to prevent the hour hand from having 
too much shake. 

As there is seldom height enough for the minute 
wheel stud to have a shoulder — which it should have 
where practicable — the plate should not be turned, but a 
small boss be left on the bottom of the wlieel to prevent 
all of it from touching the plate. 


The hour wheel should be broached to the required 
aize, and the cannon pinion fitted to it before the wheel 
is got true on the sides and finished, as, if the wheel is 
opened to fit the pinion, it fits badly and it is scarcely 
possible to keep it true. The body of the cannon pinion 
is usually left too large, necessitating a large centre to 
the hour hand. Tliis is unsightly, and has been com- 
plained of as causing a want of symmetry in the appear- 
ance of English watches. The minute wheel stud and 
pinion, and the boss of the hour wheel should be freed 
from the dial before they are finiHhed, for nothing could 
be worse than the prevailing practice of thinning and 
finishing the motion wheels before they are let into 
their places, and leaving the freeing and fitting to be 
done afterwards. 

When the third and fourth wheels have been pivoted, 
before sending the frame to be jewelled, a circle, the size 
of the balance, should be marked on the top plate, as a 
guide to the jeweller, as in case the balance comes near the 
fourth hole, the jewelling must be kept small; otherwise, 
in turning the edge of the plate to free the balance, the 
setting will be cut into. It is sometimes better to pivot 
the fourth pinion a good way into the plate, and obtain 
freedom for the balance, by running it under the jewel 
hole. All this is of course avoided in a movement of 
a correct caliber, or in a half-plate movement where the 
fourth wheel is run in a cock under the balance, this 
arrangement securing the advantage of permitting the 
fourth wheel being planted so that a larger seconds piece 
could be got in the dial. Since, in this case the hole in the 
cock is directly under the balance, it should have an 
endstone, to prevent particles or small hairs that may be 
attracted by the oil in the hole from coming in contact 
with the balance ; when the endstone is used, the pivot 
, should be conical, and may therefore be made smaller ; 
but I see no reason why this hole should not always have 
an endstone where there is room for it. 

69. Screws. — A finibher must leel that he is not 


making the best use of his time in polishing and blueing 
screws ; they are, consequently, very badly done (a lad or 
girl accustomed to the work would do them better and 
much quicker). There is, besides, seldom as much care 
taken in hardening and tempering them as there should be ; 
the screws are made too hot before hardening, and a scale 
brought on the steel that is not always removed before 
blueing them; in the case of jewel screws, the scale thus 
left on the^ is frequently fatal to the taps in the holes. 
The large screws should be hardened in a small copper box, 
or in anything that is not large and heavy and will not pre- 
vent them from cooling quickly, and that will exclude the 
air. If they are hardened at the right temperature, there 
will be no scale on them and no necessity for polishing 
the taps ; the temper should be drawn in oil. The jewel 
screws are too small for this treatment ; if a piece of 
brass wire is hammered flat and left the thickness of the 
length of the screw taps, and a number of holes be 
drilled and tapped in it as close together as the heads of the 
screws will permit, half a dozen of them can be screwed 
into the wire and hardened at the same time, and after- 
wards tempered by blazing in oil before removing them 
from the wire ; and, if the threads have been covered during 
the process of hardening, they will be quite free from scale 
and the screws be ready for polishing. If any scale is left 
on the large ncrews it should be removed by gripping the 
screw head in the Swiss screw tool, and running a thin 
screw head-slitting file through the bottom of the thread. 
A very good way of polishing the taps of screws for best 
work, is to split a piece of soft wood, making it into claws, 
charge the slit with red stuff, and by placing the screw in 
it and pressing it together in the vice, the screw can be 
polished very quickly by working it backwards and for- 
wards with a screw-driver. 

70. Suggestions as to Finishing. — I think at 
this stage the finisher's work should be completed, 
and that making and planting the stud and index, 
Bprinijing and examining, and whatever else remains 


to be done to the watch, should constitute a separate 
branch of watchmaking. It would form an intelligible 
and consistent branch, would not be too long, and time 
would not be lost, as it is with the finisher, in waiting 
for the engraving and gilding of one watch at a tima 
This arrangement would also do away with the trouble and 
annoyance caused to the springer when he finds the room 
prescribed for the spring too much or too little, the 
index pins drilled in the wrong place, and the hole in the 
stud in such a position that it must be broached (and 
to broach it now he must soften it, or adopt some more 
objectionable means of getting the stud hole for the spring 
concentric with the curb pins). Index making is already 
a separate branch, and studs are shaped and made ready 
for planting by what may be termed machinery, but 
even should the examiner, or whatever name the operator 
may go by, prefer making them himself, there would be 
a great saving in making a dozen or more at a time, and 
keeping them, ready for planting, as is now done by 
some springers of the best watches ; and making, or at 
least finishing, the iudex would be, logically, part of the 
springer's work. 

K the wheels were run in and finished, the barrel and 
fusee work completed, and the stop work and maintaining- 
power spring and detent done correctly at once, they 
would not require to be done over again ; and if the 
motion work were planted and finished and free of the 
dial, there would be little of the ordinary examining to 
do, so that fitting the watch in the case, fitting the hands, 
the stud and index work, and springing and examining, 
would form an important branch of watchmaking. The 
practice and experience gained would enable the work- 
man to do the work very much better than it has been 
done under the system hitherto practised, and to do it 
quicker and consequently cheaper. The watch could be 
sprung and examined before it was gilt, and the index, 
stud, &c., finished while the engraving and gilding were 
beiug done. There is an obvious advantage in examining 


the watch before it is gilt, as gilding has the effect 
of covering many defects of workmanship, and the great- 
est advantage this system would have over the usual 
one would be that of having one person responsible for 
the correctness of the work throughout. 

71. Applying the balance Spring.— But as I am 
not very sanguine that this suggestion will be imme- 
diately adopted or acted upon, I will proceed to 
describe what I consider the best method of applying 
the ordinary spring. The size of the spring is the first 
consideration, and this should not be left to accident ; 
it should depend on the length required, which depends 
in its turn on the closeness of the coils of the spring. 
As a general rule, half the diameter of the balance is 
a good size for the spring, but if the coils are very close, 
it should be a little less. 

If the diameter is marked on the projecting ear of 
the balance cock that carries the stud, and a notch is cut 
in it for the spur of the stud, three-fourths of it outside 
the mark, the spur may be fitted. It should be carefully 
fitted, without shake, and exactly opposite the balance 
hole ; if it is either too far back or too far forward, the 
hole for the spring will stand at an angle across it. The 
stud may be fixed with either a conical or square-headed 
screw, and it should be cut away to allow of its removal 
from the cock without taking the screw out, and the 
screw should be left long enough to permit of this 
being done easily. 

When the stud is screwed into its place, and the end 
of the spur freed from the arms of the balance, the 
circle fixing the size of the spring can be marked on it. 
If the spring collet has been made by the escapement- 
maker, as it should be, a pivot broach can be put into 
the hole in it for the spring, and the broach, lying 
parallel to the balance cock, will indicate the height on 
the spur where the hole must be drilled ; the hole should 
be drilled with a small drill. As the spring is not 
circular but spiral, the holes for the curb pins must be 


drilled inside the circle drawn for the hole in the stud, 
but how much, depends on the closeness of the coils of 
the spring, since the farther apart the coils are the more 
the spirals diverge from the circle, and care should be 
taken not to place the curb pins too far out, which is 
often the case, and is a much worse fault than having 
them too near the centre. 

When the hole in the stud is drilled and broached 
tangential to the spring, if the depthing tool is adjusted 
so that the point of the centre will mark the inside edge 
of the hole, and a mark is made on the index at the same 
distance, the curb pins can be drilled on each side of this 

A surer way than this, although giving a little more 
trouble at first, is to pin a spring in the stud, fii-st 
moving the spring through the stud hole until the proper 
size of the spring is indicated by its centre coming over 
the balance hole ; it will be seen if the stud hole is in 
the right direction, and if not, it can be broached until 
it is so at this stage with perfect accuracy. 

When the eye of the spring is concentric with the 
balance hole, if the index is placed in the middle of the 
balance cock, a mark can be made with a very fine point 
oa each side of the outer coil of the spring for the index 
pins. These should be very small, especially the inside 
one, and, if the spring has close coils, a portion of its 
thickness should be filed away, to free the second coil 
of the spring; bending the spring for this purpose 
should not be resorted to if it can be avoided. The pins 
should be close together, allowing the spring freedom, 
but as little play as possible, and only just free of the 
balance arms to prevent the second coil of the spriijg 
from getting between them. 

The horn of the index should be at such angle that, 
when the index is pushed over to " slow," it will come 
close to the stud, as the shorter the spring is between the 
stud and the index pins, the better the watch will go. 
If the spring is to be a Breguet, the overcoil will be a 


segment of a circle, and the pm holes can be marked for 
with certainty. As the spring occupies only half the hole 
in the stud, three-fourths of the hole should be outside 
the point of the depthing tool when it is adjusted to 
mark the place the spring will occupy on the index, and 
the holes can be drilled on each side of the mark. 

It is well to have a proper freedom for the spring 
between the pins without bending them, and, as the wire 
of which Breguet spiings are made is thicker than 
that used for flat springs, the holes should be drilled a 
little farther apart. If the stud of a Breguet sprung 
watch is not perfectly immovable, and the spring free 
between the curb pins in every position of the index, no 
good rate can be obtained from the watch, and there 
will be no certainty of the effect which moving the index 
will have upon its time. 

It has become the custom of the Swiss to put 
Breguet springs to watches that are quite unworthy 
of them, where no adjustment of any kind is attempted, 
and the studs and indexes are so badly fitted that these 
watches must give the worst results to the wearers of 
them and the greatest trouble to the watch jobbers ; the 
studs are made, for the most part, in the shape used in the 

best English watches. Now if Breguet 
springs are to be applied to these watches 
in this cheap and slovenly manner, the old- 
Fig. 8.— Old- fashioned Swiss stud (Fig. 8), that slipped 
Stud. ^ * into a notch in the cock, with a head on 
it, and was kept in its place by a small cap 
fixed with two screws to the cock, would be a great im- 
provement on the imitation English one. Indeed, this 
stud is a very good one, as, when the watch is set going, 
the play of the spring between the index pins fixes 
the position of the stud, when all that is necessary is 
to tighten the screws, and the spring will be in the 
middle of the index pins. 

ChapuYJn.l 103 



72. Watch Examining. — ^Watch examining has been 
a glowing branch of watch manufacturing, but with a 
better system a change should be brought about in this 
department. Competition in price compelled movement 
makers to do as little as possible to the movements, and 
the same cause and the fact of his branch being too long, 
prevented the finisher from doing a good deal of the 
work at the proper time, so that putting things right 
that should have been right at first, but were only half 
done, — in fact, doing a lai-ge portion of the work over again 
by the examiner — came to cost from a third to half the 
price of finishing the watch. But although a great deal 
of the work done by the examiner ought not to be done 
by him, there must still be careful examination and, 
when necessary, correction of every part of a watch, 
before it is placed in the hands of the wearer ; however, 
although a rearrangement of the work is desirable, I 
must treat it here as I find it. 

If the watch to be examined has a dome case, and is 
fitted to it as directed on page 109, the examiner has nothing 
more to do with that, as the fitting is complete. But 
most English watches have the double bottom case, and 
are fixed in with a bolt and joint, the bolt and joint, with 
the motion work, having been made by the motion 
maker, and the bolt polished, &c., by the finisher. As the 
finisher has probably had no opportunity of trying the 
movement in the case, the bolt may not shut in properly, 
or it may be easily released ; it must therefore be seen to, 
and the joint should be broached out with the bizzle on, and 
a pin fitted tightly enough to keep the bizzle and move* 
ment from dropping back if either or both are raised 


from the case. The pillar plate should be laid on a piece 
of plate-glass or other flat surface, to ascertain if it has 
been bent in the gilding ; if it has, it must be got flat, 
otherwise the pillars will not be upright, and will stick in 
the holes in the top plate. If the shakes are found to 
be excessive, this may be due to the gilding having thrown 
up an edge on the pillars, which prevents the top plate 
from going down to its proper bearings. In all the best 
work the finisher tits the cannon pinion to the hour 
wheel, but this is not the general practice, which is to 
just let the hour wheel on to the extreme end of the 
pinion, and not to open the pinion to the set-hand piece, 
leaving the fitting of these pieces to the examiner. The 
set-hand piece should be reduced until it fits the centre 
pinion easily ; and before the cannon pinion is let on to 
its place, it should have the square sink turned out to 
free the centre wheel stopping, when it should be 
broached to fit the set-hand piece tightly when in its 

73. Pitting the Hand& — If the body of the cannon 
pinion will not bear turning in fitting it to the hour wheel, 
the hour wheel should be opened in the mandrel, as it 
cannot be kept true by opening the hole in the fingers. 
Fitting the hands to a watch deserves more care and at- 
tention than are generally given to it. The way hands are 
fitted to English watches is bad in principle. The pipe of 
the hour wheel is left too long, and that of the minute 
hand too short, and when the end-shake of the horn* hand 
is adjusted, as it usually is, by the boss on the hour wheel 
and the dial, the end-shake of the centre wheel affects it, 
sometimes giving it too much and bending the hour 
hand by its catching the minute hand either in setting 
the hands or in the going of the watch. In fitting the 
liands, the examiaer should fit the glass, if to a hunting 
case, as high as the case will admit, ascertain the space 
available by placing a piece of bees-wax on the dial and 
pressing the glass down on it, and turn the cannon pinion 
until it projects from the dial the height of the bees- wax ; 

Cbap.Yin.] WATCH EXAMINIXO. 105 

the lionr wheel pipe should rise just j)erceptibly above the 
dial, and the end-shake of the hour hand be adjusted hy 
the pipe of the minute hand and that of the hour wheel. 

If done in this way, the end-shake of the centre wheel 
will not affect that of the hour hand, which will be always 
the same, and may be very little, and the minute hand 
will have a sufficiently long and secure fitting ; the hour 
wheel pipe is always long enough to make the fitting of 
the hour hand perfectly safe, except where a very thin 
gold dial is used. The haphazard method of fitting the 
hands without measuring the room in the case is a very 
bad one ; they are generally fitted too low, and, with a 
badly fitting hour wheel, are constantly catching, and so 
stopping the watch. 

As with the hour wheel, the pinion should also be 
fitted to the minute hand, and not the hand opened to 
fit the pinion ; when the hand is fitted, a little red stuif 
can be placed on the ball, and the pinion lowered until 
the red stuff is quite free of the glass, when the end-shake 
of the centre wheel is pushed towards it ; the ball of the 
hands and the ends of the pinion and set-hand piece are 
then ready for polishing : this part of the work is usually 
well done. A very convenient tool for doing it is the 
Swiss screw-head tooL 

74. Attaching the Dial. — The pin holes in the dial 
feet should be drilled with a very small drill, in such 
a direction that the pins will not come in the way of any- 
thing, and will be easily got at ; they should not be drilled 
below the surface of the plate, but broached until the pin 
touches it. If the hole should be a little below the sur- 
face, it is better to lengthen the copper foot by squeezing 
it with a pair of blunt nippers until it is above the plate 
than to leave it in such a position that no pin can stop in. 

There could not be anything much worse than the 
manner in which the dials have been attached to common 
English watches, the holes for the dial feet being placed 
in the most awkward positions, the pin holes drilled so 
very much too large that they were constantly bursting 


out at one side, and so far under the plate, that the pins 
must be bent to go into them and pushed in so tightly 
to prevent them from falling out, that the dials were and 
are frequently broken by the mere process of pinning them 
on after cleaning the watch. There is no better way 
of attaching a dial to a watch than by pins through the 
feet if it is properly done ; but two dial feet are better 
than three, if one of the three is in such a position that 
a pin cannot be got into it. 

75. General Revision. — The mainspring should be 
removed from the barrel and the barrel freed on its 
arbor, if it requires a little end-shake, as it usually does. 
The freedom should be taken from whichever side of the 
barrel is nearest the frame : i.e., if the barrel should be 
nearer the top than the pillar plate, the boss of the cover 
should be tliinned to get the inside freedom, and vice versa; 
the mainspiing can then be put in and adjusted. With 
the adjusting rod on the fusee square, any fault in the 
action of the stop work will be seen, and it will also be 
seen whether or not the chain leads properly on to the 
fusee. The spring should be set up two or three teeth be- 
yond where it makes an adjustment, as it will give a little, 
and the set-up be marked on the plate and barrel arbor by 
a couple of small dots. When the fusee is taken out of the 
frame, the adjusting rod should be put on the square, and 
held horizontally with the great wheel in the left hand ; 
if the weight of the rod, which is a measure of the 
strength of the mainspring, draws the maintaining spring 
to the end of the notch cut in the great wheel, the main- 
taining spring will be weak enough, but if not, it must 
be weaJjened until it does so. The maintaining power 
and great wheels must not be pinned on too tightly 
to the fusee, for if this be so, or if the maintaining spring be 
too strong, the maintaining power will not act ; and the 
examiner should see to this when adjusting the fusee, 
and also see that the stop acts properly. (See p. 93.) 
The holes for the set-hand and winding squares must be 
made in the case, or, if they have been drilled, opened to 


the required size, and the squares reduced to the proper 
length. Polishing the ends of the squares was at one time 
thought to be of importance, as it gave an appearance of 
finish to the watch, and often veiy bad watches had the 
ends of the squares polished flat ; there is, however^ no 
necessity for doing this even in the best work. If the end 
of the winding square is left flat, it is difficult to put a key 
on that fits it, and it soon gets scratched ; nor should it 
be rounded, like some of the squares to the best Swiss 
watches, since that makes the square too short. The 
square should be shortened gradually, until a little red 
stuff on its end does not touch the bottom of the case 
(the set-hand square may be a little lower), and should be 
polished either in an English or Swiss screw-head tool ; 
the latter is more convenient for this purpose, especially 
for the set-hand square ; the end should be just a little 
rounded and the narrow comers be taken off with a 
very old and smooth pivot file. 

The wheels should be put into the frame separately 
and the end and side shakes of each noted ; the depths of 
the great wheel and centre pinion, and centre pinion and 
third wheel can be seen without putting on the top 
plate ; the fourth and escape depths must be tried with 
the frame together, as they are more important and less 
easily seen, and, if doubts be entertained as to their 
correctness, they should be put in the depthing tool and 
examined. In going barrel watches, the examiner has 
less work ; he should see that the spring is properly free 
in the barrel, and of a proper length, and not set up so 
far that it is in danger of pulling the hook away every 
time it is wound up ; the set-square must be kept quite 
free of the bottom, and it should be fitted very well to 
the centre pinion, but loose, and, before putting the 
watch together, a slight dent should be made in the stem 
with a sharp-pointed punch ; this will give it a kind of 
spiing tightness in the pinion which will permit of the 
hands being easily set and at the same time keep the 
minute hand sufficiently rigid. 


If the stem is fitted tightly, by polishing, into the 
pinion to keep the minute hand from being too easily 
moved, there will be a great danger of it sticking fast ; 
it should always have a little oil put to it to prevent 
what- is termed firing. It will save a little labour, and 
be a good deal better in principle, to revert to the old plan 
of a solid centre pinion for keyless watches ; by snapping 
or springing the cannon pinion on the solid arbor, setting 
the hands is much more easily accomplished, and there is 
another advantage in having the centre pinion's pivots 
smaUer and stronger. 

76. Fixing Movement in Case.— As the largest por- 
tion of English watches are made with what are called 
"double-bottom" cases, i.e., cases having the inner 
cover or bottom made solid with the middle, the only 
way of fixing in the movement that will permit of it 
being opened to view the works is with a joint and bolt ; 
but there are so many objections to this arrangement 
that it is difficult to account for its continuance. In the 
first place the edge of the pillar plat« of the watch is cut 
through, which weakens it so much that it often gets 
bent in the gilding ; then, the glass bizzle must be 
opened before the movement can be seen, and, as the 
bizzle opens at right angles to the cover in a hunting 
watch it always marks it. The necessity of making the 
bizzle open easily, and making a joint on it, prevents it 
from fitting closely, and from excluding dust and damp 
from the dial and movement ; and there is also the danger 
of an inexperienced person pulling ofi" the hands and 
breaking the dial in opening this bizzle and lifting the 
movement from the case. 

The English case-makers are principally to blame for 
the preponderance of double-bottom cases, for, being 
unaccustomed to make dome cases, they charge one- 
fourth more for the fashion (i.e., the making) of them 
than for the making of double-bottom cases ; the Swiss, 
on the contrary, reverse this order, and charge more for 
the latter. 


The case-makei's muut not, however, have all the 
blame, since, had the manufacturers chosen, they might 
have altered the system, as the price of the bolt and 
joint would have gone some way to pay the extra cost of 
the dome case. But other reasons may be found in the 
fact that a dome case requires more gold in the middle 
and bands, and the foolish law that prevents the English 
watchmaker from fixing, even by a pin, a dome to a 
case of any other metal than gold. Thus law and custom 
have prevailed against common sense and economy, and 
I believe that full plate watches, with double-bottom cases, 
are responsible to some extent for the decline of watch- 
making in England. It is therefore to be hoped that 
some means may speedily be found of superseding both, 
although habit has so much influence that when dome 
cases are used the movement is often fitted to the case 
with a pin in the edge of the plate and a small hole in 
the projecting band of the case, and a joint and pin on 
the opposite side ; even keyless watches being made with 
a bolt and joint. It is not usual to let the finisher have 
the case, as it can be sprung and polished, &c., while 
the watch is being finished, and, if the watch has a dome 
case, it is necessary to fit the movement to it first. The 
best way to Ek it is with one pin and one screw ; two 
pins and a screw are sometimes used, and this may be 
done if the edge of the case where the pins go is deep 
enough to take pins of suflficient strength, but this is 
seldom the case. A good-sized hole is drilled in the edge 
of the pillar plate close to the step, and a piece of brass 
is tapped and screwed into it, projecting a very little 
beyond the edge of the plate ; this piece is filed and the 
sides of it shaped with a chisel into the form of a wedge, 
when the position of the dial in the case is marked on 
the flange of the case which supports the frame, and a 
square notch is cut in it to receive the thick part of the 
pin, and prevent the frame from moving round in the 
case, while the wedge-shaped end of the pin projects 
underneath it, and prevents the frame from rising from 


the case. A mark is then made for a screw on the top 
plate opposite the wedge, and drilled close enough to the 
edge of the plate for the head of the screw to come over 
the band of the case, and a dog-screw (i.e., a screw with 
a portion of the head cut away) fitted; this screw is 
screwed into the pillar plate, the body going through a 
hole in the top plate into which its head is sunk as deeply 
as the thickness of the band of the case will admit of. The 
sink in the plate must be a very little below that in the 
band, and the screw should have a shoulder to rest on 
the pillar plate while the head grips the band of the case 
sufficiently tightly to prevent any movement of the 
frame in the case. I have found this the most con- 
venient and simple way of fitting a movement to a case, 
as, if it is properly done, the movement can be removed or 
put in the case with the greatest ease. In putting it 
into the case, if the movement is moved round the 
wedge will drop into the notch, and, by turning the 
screw half round, the movement is secured. Care must 
be taken to find a place for the screw that will be free 
from the other parts of the watch ; in ordinary fusee 
movements, between the figures seven and eight on the 
dial will be the best place for the pin, and there will be 
room for the screw opposite. 



77. Objections to Jewelling. — Jewelling the holes for 

watch pivots has been practised for nearly two centuries ; 
it was invented by Nicholas Facio, a native of Geneva. 
He came to London in 1700, and a few years later com- 
menced the business of watchmaking and watch jewelling 
in partnership with the brothers De Beaufr^. The 
watchmakers of Paris, to whom he first applied not 


appreciating his invention, gave him no encoumgement, 
and the London watchmakers do not appear to have 
treated him with much greater liberality, as the Clock- 
makers' Company opposed his application for a patent, 
although, presumably on the strength of his invention, 
he had been admitted as a member of the Royal Society. 
Watchmakers of a century and a half ago were con- 
sidered of more importance than they are at present, 
and many of them were members of that learned 

Notwithstanding any jealousy the watchmakers had 
of Facio, his invention was speedily adopted both here 
and on the Continent. The jewelling was for the most 
part confined to the holes of the escapement, probably in 
consequence of its cost, but there must have been a 
prejudice against jewelling watches at an early date, 
which has not yet quite disappeared, because of the 
tendency of certain sorts of stones to blacken the pivots 
when the watch has gone for some time and the oil 
turned viscid. However, there are no watches made now 
that have not at least the balance staff holes jewelled, and 
I do not think there are any watchmakers that will not 
admit that the holes of the escapement would in all cases 
be better for being jewelled; although I have heard 
men argue that the third and fourth wheel holes of a watch 
would be better to be of good brass than to be jewelled. 
I am sure there are thousands of English watches that 
would have gone longer and cost less to repair, if they 
had been jewelled in a few more holes. The Swiss, on 
the other hand, jewel their very worst watches in as 
many holes as possible, and do it so badly that brass holes 
would in many cases be preferable. 

78. Stones for Jewelling. — The stones used for 
jewelling watches are the ruby, sapphire, chrysolite, and 
garnet ; a thin rose diamond is generally put as an end- 
stone to the balance cock of English watches, but only as 
an ornament, and that is the only diamond ever used in 
the jewelling of a watch. There is an uncharitable 


belief that watchmakei'S sometimes change the jewels in 
watches for stones of inferior value, but there is no founda- 
tion for the calumny, and the time spent in making the ex- 
change would certainly exceed the value of the best holes. 

There are great varieties of all these stones, so that 
it cannot be said that a ruby is best for a hole unless it 
is the right sort of ruby, and colour is not always a guide 
as to the quality ; the oriental ruby is the best, being the 
hardest and having the greatest specific gravity; it 
should always be used for the best watches. Sapphire is 
usually used for the holes of marine chronometers. 
Rubies that have a deep red colour are prized and used 
by the Swiss, while in England the milky stone is pre- 
ferred as being harder ; it is also thought that, in con- 
sequence of the colouring matter, the red stones blacken 
the pivots more than the light-coloured ones. 

My own experience is that, if the stone is hard and 
well polished, the colour is not of much consequence ; 
since, although the pivot becomes black, it does not cut, 
and the discoloration is easily removed with a peg and a 
little fine red stuJQT. 

The quality of the oil has much to do with the blacken- 
ing of the pivots, and those which have the greatest fric- 
tion will become discoloured first. In ordinary watches 
jewelled in the third and fourth wheel holes, the lower 
third wheel pivot will be the blackest, it having the 
greatest friction, from being so close to the action of the 
centre wheel in the pinion ; and if the centre holes be 
jewelled, the bottom pivot will generally be found more 
discoloured than the top one from the same causa But 
there are so many reasons in favour of good jewel holes 
that every good watch should have all the train holes 
jewelled except those of the fusee, which are expensive, 
and liable to be broken with the pressure in winding the 
watch. Garnet is largely used for jewelling common 
watches, especially in the pallets to lever escapements ; 
it is of the same hardness as chrysolite, but not so brittle. 
These pallets are soon cut, a few years' wear pitting the 


rubbing face of the stone on which the escape- wheel tooth 
drops, in which case the only remedy is new pallets, as to 
polish out the pits would spoil the escapement. Chryso- 
lite would answer better for pallet stones, only it is not 
so like ruby as garnet is. Garnet is also used for the 
impulse pins of lever escapements, but the least violent 
external motion to the watch will break off the pin, if 
the balance be a heavy one, and the cost of replacing it 
will be many times the difference between the original 
price of a ruby and a garnet pin. ^ 

The test of the hardness of pi'ecious stones is in the 
working of them; their specific gravity can be ascertained 
by weighing them in tiie air and when suspended in 
water, and by dividing their weight in the air by the 
difference ; the water for this purpose must be distilled. 

Watch jewelling in England has hitherto been divided 
into two bitinches only ; namely, hole-making, and jewel- 
ling ; the hole-maker flatting and drilling the stone, and 
turning it true on the edge and faces ; and the jeweller 
fitting the hole to the pivot^ shaping and polishing it, 
and setting it in the plate. In Switzerland there is a 
great division of labour in preparing and fitting the holes 
to watches. Jewelling would seem to be as far removed 
from escapement-making as two branches of a trade 
could be, but the Swiss escapement-maker is served with 
the jewel holes along with the other materials for the 
escapement, and sets them himself. This practice would 
be useful to watch jobbers going to India or the colonies, 
but watch jewelling is so much of a specialty, and 
requires such great practice to do it well, that one is 
surprised at the combination of two such parts surviving 
in Switzerland. 

79. Hole Making. — The English process of making a 
hole is to take the rough stone, about the size of a small 
pea, and hold it against the face of a plate or mill fixed 
in the lathe and rotated rapidly ; the mill is of soft iron 
charged with diamond powder. When one side of the 
stone is flatted, the other side is held against the mill 


until the stone is brought to the required thickness ; it is 
then cemented on to a chuck, turned true on the face and 
edge with a piece of black diamond fixed in a handle, and 
centred with a small splint of the same, and drilled to half 
the length of the hole ; the stone is then reversed on the 
chuck, the face turned true, and, if it is the front of the 
hole, the chamfer or cup is turned out of the centre, and 
the hole met. From first to last this seems a very slow 
process ; and flatting the stones is a very dirty one, as 
the mill must be kept supplied with plenty of water. 
In fact, I believe it is the identical method pursued by 
Facio, the inventor. When the hole is a large one, the 
process of drilling with a diamond point is well enough, 
and it is acknowledged that holes made in this way 
cannot be beaten ; but, without adopting all the Swiss 
system, which has some drawbacks, I think ours ought to 
be greatly improved. 

The Swiss flatten the stones on a large horizontal 
mill driven at a very high rate of speed, generally by a 
turbine, or by steam power. The stones are not presented 
singly to the mill, but are cemented on to a block, and 
held against it in quantities of some dozens at a time ; 
when the stones are sufficiently reduced on one side, they 
are reversed, and the other side ground until they are the 
required thickoess. It is evident that this operation can 
be done for a tithe of what it would cost to have them 
flattened by the old process. 

The uncertainty of getting the sides parallel to one 
another by the above method, however, prevents the holes 
from being drilled perpendicular to one of these sides. 
If they were drilled one at a time this could be d(xie, but 
the stones are pushed half a dozen or more into a kind 
of tube or holder, which is fixed on the rest of the lathe 
in a line with the drill in the chuck. This drill, instead 
of being a diamond, is a piece of drawn steel wire ; it 
goes through a great many stones at one time, and is 
charged with diamond powder, and, instead of a man or 
woman drilling one hole and then stopping the machine, 


he or she has to attend to six^ machines, each drilling six 
holes at one operation, and the faces of the holes not 
being parallel to one another, the holes are seldom 
perpendicular to either side of the stones. If they are 
much out it is not possible by any process of opening to 
make true holes, especially if they are a good length ; but 
this process of flatting and drilling is so expeditious, and 
consequently cheap, that jewelling an extra pair of holes 
makes little or no difference in the price of a Swiss watch, 
and jewelling every hole is now the rule with them, 
although, as before stated, they are not always as good as 
brass ones. They are made very thin, are very badly 
polished, and of such material that they are so easily 
broken that few common Swiss watches are without a 
cracked jewel hole or two. Some years ago there was 
quite an industry in London of finishing these holes for 
America, — so it was believed, — but now I understand the 
factories in the States make their own holes, and there is 
certainly no reason why we should not take a leaf out of 
the Swiss book, and amend our system without alto- 
gether adopting theirs. A Swiss jewel hole maker 
informs me that thirty years ago, before the drilling in 
quantities was adopted, a woman or lad coidd drill a 
himdred flatted stones in a day, and the holes in these 
stones were required to be perfectly perpendicular to one 
side of the stone, although the rate of speed that a boy or 
woman could drive a foot-lathe would be much too slow 
for a revolving driU ; but this process would seem a great 
improvement on ours of setting up a stone in a lathe, and 
meeting a hole small enough for a staff pivot. 

80. Jewelling. — When the jeweller first gets the 
frame to jewel the holes for the escapement, he has 
nothing to do with fitting the pivots, as they are not then 
made, but he should have a size given him, and his care 
should be to make use of all the thickness of the plates 
for holes with end-stones, to make the sinks in which the 
settings rest quite square, and the shoulders of the set- 
tings to coincide with the sinks. The setting itself should 


have some substance in it, and not be turned away so thin 
that any extra pressure on the end-stone would bend it in 
and reduce the end-shake, or until it is impossible to lift 
the hole, unless it is done as the jeweller does it, on the 
end of a damp finger. Some watchmakers object to 
large stones on the ground that the colouring matter 
being in a greater body than in a smaQ one, it wiU 
blacken the pivot more ; but then, on the other hand, it 
should be remembered that a small stone means 
necessarily a weak one, and a large stone has other 
good qualities as well as strength. When there are end- 
stones, the oil chamfer is cut at such an angle in a thick 
stone that the pivots drop into their places without 
trouble, and without the risk of injuring them in putting 
the watch together — a thing that is so often happening to 
the pivots of the balance staff, especially where the 
balance is a heavy one. Some jewellers make a double 
chamfer to the holes, but I see no use in that. 

There has been considerable difference of opinion 
amongst watchmakers as to the best shape of a hole : 
some have advocated a long straight hole with a pivot, 
largest at the extreme end to lighten the friction (as 
shown in Fig. 28) ; but no person who has had much ex- 
perience of the going of watches would think of making a 
balance-staff pivot unnecessarily weak, and of the very 
form most liable to injury. A jewel hole should not 

be straight, but rounded from both ends 
to the middle, so that the rubbing surface 
Fig. 9. shall be small and equal, whatever the 

amount of end-shake maybe; as shown in 
Fig. 9. This is also the best shape for thorough holes (i.e., 
holes without end-stones), although they are seldom made 
so. English jewellers have persistently used screws with 
too small heads and too large taps for fixing jewel holes 
with endstones. There is no advantage in a large screw 
wheriB there are only three or four turns of thread on it, 
and jewel screws seldom get broken ; occasionally half 
the head comes off if it has been slit down too much. 


The objections to large jewel screws are that, in order 
that the heads may come over the setting, the holes 
have to be drilled so close to the sink in the plate that 
in many instances they burst out, and if, from overheating 
the sci*ew in hardening, a scale is left, or any other 
accident happens from carelessness on the part of the 
examiner, &c,, and the thread of the hole is injured, the 
hole cannot be tapped over again with a larger tap, as 
could be done should this occur with a small hole. 
I have seen many watches almost new, that purported 
to be good ones, with several of the jewel screws over- 
turned and quite useless, and this must have been the 
case before they left the maker's hands. This is especially 
the case with the jewelling of escape-cocks that have 
end-stones. Unless the cocks have ample thickness, and 
can be properly jewelled, they would in most instances 
be much better with thorough holes. Jewellers can 
hardly be blamed for doing the work in this way ; they 
work for watchmakers who ought to know more about 
the matter than they do. The bodies or taps of the 
screws should be small, and the head large enough to 
come over the setting, keeping the body of the screw 
at such a distance from the sink that, if the screw should 
be overturned, the hole could be re-tapped, and a larger 
screw put into it without bursting the hole into the sink. 
The bottom holes of the third and fourth wheels are 
rubbed into the bar, and the holes in the top plate are set 
in gold or brass settings, which are countersunk into the 
plate and fixed with two screws, the heads of which are 
sunk partly into the plate, and partly into the jewel 
hole setting. In many of the watches made by the 
Roskells, and other Liverpool makers, the holes were 
all rubbed into the plates, and they had thorough holes 
to the escape wheel and pallets. This was an excellent 
plan, as many of these watches went to South America, 
and to other parts, where the watchmakers were not very 
skilled workmen, and, as the holes were all set in this 
manner, they had no trouble with the jewel screws, and 


I have seen these watches after many years' wear very 
little the worse for the jobbers. 

It is now, however, the fashion to put end-stones to 
the escape-cocks, very often where there is insufficient 
room for them (and where, in consequence of the want 
of room, the work would be unsound, even if well done) 
because end-stones are characteristic of good work, and 
one sometimes sees an end-stone to the escape- wheel, with 
the pallet staff in a brass hole. First-class jewelling 
costs nearly as much in Geneva as it does in London, 
notwithstanding the superior methods the Swiss have of 
flatting and drilling the stones, as opening, setting, and 
polishing the holes require much greater skill and labour 
than making them. The holes are opened with copper 
wire and diamond powder, and polished with diamond 
powder on a hard dog-wood peg, and a hole requires a 
good deal of labour with the latter before it is properly 
polished ; but, notwithstanding this, if English watch- 
makers would set about reforming our present system 
by introducing a greater division of labour, and the Swiss 
Bystem, in part, of making the holes, sound second-class 
jewelling might be obtained at much less cost and of a 
superior quality to what is at present to be got. 

81. Snailing. — ^There is the objection to polishing 
brass surfaces such as great wheels and barrel covers, 
that, although they are somewhat difficult to polish, they 
are easily scratched, and that brushing spoils their appear- 
ance. The art of spotting such small pieces by hand is 
not easily acquired, and gilding great wheels and barrels 
with teeth on them has proved so ruinous to the work by 
softening the brass that it is rarely resorted to now. 
Snarling is a very old method of finishing the steel caps 
of fusees, but since the introduction of keyless work it 
has come more into use, it being the usual way of 
finishing the steel winding wheels. The tool usually 
employed by finishers by which they perform the 
operation is a very primitive one, mostly consisting of a 
copper penny driven on to a good-sized arbor, and 

Chap. IX.] SNAILINa. 119 

termed a mill. The copper is turned away from the side 
of the roller that is to be used until a thin rim only is 
left projecting at its outer edge ; the roller is fixed in a 
pair of small turns which are fixed to the rest holder of 
a larger pair by a projecting shank ; the work that is to 
be snailed is put into the large turns, and the snailing 
roller brought close to it. One of the runners of the 
small turns is excentric to the other, and by turning this 
runner round, the roller is brought into contact with the 
top side only of the wheel to be snailed ; if the faces were 
parallel to each other, the curves made by the roller 
coming in contact with the work would be crossed and 
obliterated by the roller when leaving it, even - if the 
arbor turned only in one direction. This effect is to 
some extent produced by using a bow, as the up strokes 
of the bow make curves in a contrary direction to those 
made by the down strokes, but this confusion of circles 
does not matter very much, as when it is found that the 
roller has been in contact with every part of the work, 
and that it is smooth, the snailing is done with the down 
strokes only, a long bow being used, and the snailing 
roller held in the fingers and prevented from turning, 
while the bow is pushed upwards. Emery and oil mixed 
with sharp stuff is recommended with the copper roller, 
but copper is too soft and a bad material for a roller, as 
the emery sticks in it, and it is difficult to get a smooth 
surface with it or without deep races in it ; the roller 
should be made of hard brass, and the emery should be 
washed until it is very fine, and no sharp stuff used with 
it Brass wheels, <fec., should be stoned free from scratches, 
and got quite flat on arbors. For this purpose a roller 
made of hard box-wood should be used with oilstone dust 
ground as fine as possible, and mixed with oil to the con- 
sistency of cream, and the polishing power used sparingly. 
Sharp stuff and oil gives a bright surface on brass, but 
polishes too much, leaving the circles undefined. Boilers 
of any size may be used ; one twice the diameter of the 
object snailed makes a curve that looks yery welL 


Snailing is a very excellent way of finishing great 
and motion wheels, and the bottoms and covers of 
barrels, and if done as directed, they will brush bright 
and clean without showing any scratches. Although 
it is in universal use, the tool described is obviously 
only a makeshift, and the difficulty of using it properly 
may be judged of from the number of badly-snailed fusee 
caps one sees, the fusee cap being all that is snailed in 
ordinary watches. 

If polishing should be made a special branch of 
watch finishing, snailing might be well done without the 
application of very great skill by using a proper tool, 
adapted to either a foot-lathe or a hand-lathe, or throw, 
with a rotary motion, with a screw for the adjustment of 
the angle of the roller. As the snailing roller will not 
go right to the centre of a wheel but would leave a 
lump there, a hollow is usually cut close to the centre 
and polished ; this gives a little variety and relief to the 

82. Gilding. — The art of gilding base metal is of very 
early origin, and must have been applied to the frames 
and wheels of the first watches made. I have seen 
watches by Mudge, Graham, and other early watch- 
makers, with the lilding so fresh and perfect that, but 
for the name on them, it would not be possible to tell 
they were more than a few years old. But the process by 
which these watches were gilded has almost died out since 
the advent of electro-gilding, although it is well known 
that brass gilded in this way will last for centuries. 
Not ten per cent, of the watches made in England are 
now gilded by the mercurial process, and none of the 
Swiss ones. 

English watchmakers have been always slow to 
accept new ideas as such, and have paid less attention to 
appearance than utility ; and, as brass is a better material 
for watch frames and wheels than any other yet dis- 
covered, they have kept to brass, unless in the case of 
some special order. The Swiss, on the other hand, have 

C1»P- nL] GILDING. 121 

been more energetic and more enterprising in finding 
a market than their English competitors, and in 
seeking something novel have adopted a white metal, 
which they call nickel, for the frames and bars, and a 
low quality of red gold for the wheels. To heighten the 
effect, they employ the reddest stones they can find for 
]ewel holes, and set them in red gold, the wheels being 
polished, and the plates or bars snailed, spotted, and 
polished in a variety of ways, by means of which a very 
pleasing contrast is obtained. 

This attention to novelty and beauty of appearance 
has undoubtedly assisted the sale of Swiss watches, 
especially in the United States, where they imitate the 
Swiss style, and make all their best watches with white 
metal frames. But this so-called nickel is only German 
silver, and although it looks pretty at first and keeps its 
colour for a time, it gradually gets dim and tarnished, and 
will not last like the old-fashioned gilding, nor is it so 
suitable as brass for screw or pivot holes. 

As the Swiss no longer make their watches with 
brass frames and wheels, unless they are for the English 
market, the gilding has deteriorated there, and they now 
send their best wheels here to have them gilded. 

Mercurial gilding has a rough granulated appear- 
ance, from the quantity of gold on the surface and the 
manner in which it is laid on. When gilding by the electro 
process was introduced, it was thought desirable to 
imitate the appearance of the best gilding, but a heavy 
deposit of gold cannot be left by this process without it 
running into circles and inequalities in consequence of 
the holes in the plates, and as electro-gilding is practised 
because of its cheapness very little gold is left on, and 
the desired surface is obtained by softening the plates 
and scratch-brushing, i.e., brushing with a wire brush. 
There is no reason why electro-gilding should soften the 
brass if the smooth surface left after the process were 
not interfered with, but in any case the gold does not 
adhere to the brass so well as in mercurial gilding. 


The Swiss obtain a surface by depositing a thick layer 
of tin on the plate or other article, scratch-brushing the 
tin, and gilding afterwards, the gold being so lightly laid 
on that it would not stand much brushing after gilding ; 
but this method is very unsatisfactory, as, in cleaning 
the watch a few times, the gold is brushed off the edges 
of the plates and bars, making the work look worse than 
if it had never been gilded at all. 

The mercurial process is as follows. Take, say, half 
an ounce of fine gold, and by the application of a little 
heat, dissolve it in about three ounces of mercury ; the 
work to be gilded must be prepared by placing it for a 
short time in a solution of dilute nitric acid and 
mercury. The dissolved gold and mercury are put into 
a small wash-leather bag, and the mercury squeezed 
through the pores of the leather until the paste in the 
bag becomes of the consistency of soft butter. The article 
to be gilded is then heated a little, and the paste spread 
over it with a piece of brass used for this purpose, 
technically called a pencil. The mercury remaining in 
the paste on the plate is evaporated slowly by heating 
it ; the heat applied should never be greater than the 
temperature of boiling water, so that the brass be not 
softened ; when the mercury has all been destroyed, the 
plate is scratch-brushed, with an acid — usually beer — to 
bring up the colour. Although the process seems simple 
enough, great skill and experience are necessary in the 
operator, and there are few good mercurial gilders left 

Polishing the train wheels of watches was much 
practised in Liverpool, when watchmaking was more of 
an industry there than at present, and polished wheels 
looked very well for a time ; but, as in the case with 
polished plates which have sometimes been adopted as a 
distinguishing feature for pocket chronometers, they 
soon tarnished. 

A very pretty effect is produced by polishing and 
spotting a plate and electro-gilding it afterwards ; the 
spotting shows through the light coating of gold and 


looks very rich, and the gold prevents the plate from 

Great pains are occasionally taken to get the gold 
out of the pivot holes after gilding, but in the case of 
large holes this is a mistake : a well-gilded hole will last 
l^ger than a br^s one. and 'wiU pres^e the oil instead 
of turning it green as brass always does. The late Mr. 
Blundell told me he always got the holes of the regu- 
lators he made gilded when jewelling was too expensive. 

It is a question whether English watchmakers have 
been wise in disregarding the well-known desire all 
people have for new things, and in adhering to brass as 
the principal material for watches, especially as there are 
now so many alloys of aluminium of various colours and 
other metals that might be made available for the 
purpose, which are much better adapted to variation in 
the mode of finishing than brass is. 



83. What is an Escapement? — The escapement of 
a watch or clock is that part of the mechanism which 
controls the speed of the train of wheels, and compels 
the motive force to exhaust itself uniformly by allow- 
ing only one tooth at a time of the last wheel of the 
train to escape. Hence the term. This last wheel is 
called the escape or 'scape wheel, and is included as part 
of the escapement, every different kind of escapement 
having a peculiar escape wheel. 

The escapement, then, consists of this wheel, the 
balance, and all the intermediate pieces. The escapements 
used in watches are five in number — although many others 
have been invented from time to time which have never 
come into general use. They are the verge, the horizontal, 


the lever, the duplex, and the detached or chrono- 

84. Verge Escapement. — Although the first of these 

escapements may be considered to all practical intents 
obsolete, no satisfactory results being obtainable from it, 
yet, as it still continues to be made, it is necessary for 
me to notice it. M. Saunier says, in the second edition 
of his Treatise on "Modem Watchmaking, "There are still 
made, chiefly in the canton of Berne, more than 300,000 
verge watches annually," but I think he would now 
considerably modify that statement. 

The verge, vertical, or crown wheel, is the oldest 
known form of escapement. It was applied to the clock 
made by Henry de Wyck for Charles V., but the inventor 
of it is not known, though by some it is ascribed to Pope 
Sylvester II. 

It is a recoil escapement, entirely frictional, and has 
no free arc, the supplementary arc taking place during the 
recoil, and consists (Fig. 10) of a crown escape wheel, 
which gives impulse to the balance through two pallets 
fixed at about a right angle on its arbor, the minute 
wheel of the train being also a crown wheel. 

In Fig. 10 the diagrams A and B represent the plan 
and elevation of the escape wheel, with the portion of the 
balance staff, or verge, upon which are the pallets which 
control it. 

The wheel is travelling in the direction of the arrow, 
and the tooth under the pallet, a, is just escaping after 
having given impulse to it. When the tooth under a has 
escaped the tooth under b will come into action with that 
pallet, and, after a recoil caused by the inertia of the 
balance, will in its turn give impulse until it escapes, 
when the pallet a will receive the next tooth in like 
manner, the motion being reciprocal, the pallets acting 
alternately at opposite sides of the wheel. 

It is obvious that the escape wheel must have an odd 
number of teeth in order to'act at all. The width of the 
pallet's and the angle at which they are set to one another 



*is determined by tlie number of degrees through which 
the balance is required to move in its vibrations. 

The greater the angle at which thej are eet the 
greater will be the vibration and the less the recoil, and 
vice vered, as, of course, the greater the recoil the greater 

will be the retaidation produced upon the motion of the 

Writers on ijie subject differ as to the number of 
degrees, varying from 90° to 115°, of opening between 
the pallets proper for this escapement, ^Engligh watch- 
makers generally used the smaller openings in their 
escapements, while the French made tbem with the 
larger, pitching the pallets deep into the wheel and 
cutting away a portion of the verge to get the teeth to 
act on the pallets nearer to its centre and thus give a 
longer impulse. But these escapements, unless veiy 


carefully made, were liable to' set owing to the very small 
leverage on the tooth commencing its action of giving 
impulsa To prevent this, they were usually made with 
very light balances. They were generally superior to 
the English #escapement both in principle and make. 
With the extremely small angles the large amount of 
recoil retards the vibrations, so that practically it has 
been found best to place the limit between 95" and 105^ 

In all cases the wheel should be planted as closely as 
possible to the centre of the verge so that the points of 
the teeth shall just free it, and, that the lift may be 
equal on the pallets, should also be planted so that the 
verge shall cross exactly its centre at right angles to it. 

The pivot of the escape wheel nearest to the verge 
works in its hole in a dovetail in the potance, and in 
setting the escapement a little more space should be left 
between the top pallet and the wheel than between the 
lower one, to allow of more. drop. This is necessary, 
owing to the weight of the balance pressing that pallet 
deeper into the wheel (the distance of the shake in the 
balance cock hole), when the watch is in a vertical 
position with the verge above the wheel than when it is 
lying flat ; verge escapements very often tripping in the 
one position while they are all right in the other, where 
this precaution has not been adhered to. 

The escape wheel should have no perceptible end- 
shake or the tripping will occur, whenever the wheel is 
by the amount of the shake too deep, while on the other 
hand the vibration will fall off when it is too shallow. 

85. Horizontal Escapement. — The horizontal or 
cylinder escapement was invented by George Graham 
about the year 1700. It was the development of 
an idea conceived by Tompion, who had already in- 
vented and constructed an escapement of a similar 
principle, and may be considered to be the first step 
in the direction of good timekeeping, as regards the 
principles of an escapement for pocket watches hitherto 
made, being the first dead beat escapement applied to 


them. Its advantages, as compared with the verge, were 
obvious and numerous (even as at first constructed) 
although it did not quickly supplant it. The wrong pro- 
portions at first adopted and the comparative difficulty 
of its manufacture rendered it unsatisfactory in its 
results, and unpopular with most of the watchmakers of 
the day. 

Fig. 11, A, shows a plan of the escape wheel and 
cylinder. The wheel is travelling in the direction in- 
dicated by the arrow, and the tooth a just commencing 
to give impulse to the left hand, or entrance lip, of the 
cylinder. When it has given impulse, it will escape from 
that lip and fall into the position of the dotted tooth 
a' at Fig. 1 1, b, the right-hand edge of the cylinder being 
in the position marked by the dotted lines at 1, and con- 
tinuing to travel until it reaches the point 2, when the 
inertia thus given to the balance being overcome by its 
spring, it will return to 3, when the tooth a will give 
impulse to the right hand, or exit lip of the cylinder, 
escaping as at 6. The tooth c will then fall on the out- 
side of the cylinder, as shown, the supplementary arc 
continuing to 2 and returning, the tooth c then giving 
impulse through the left-hand edge as .before. 

Fig. 11, c, shows the cylinder in elevation, containing a 
tooth of the escape wheel, the balance being at rest, and 
the power of the mainspring off", or the watch being 
down ; m shows the space which is cut away to free the 
escape wheeL 

Supposing the balance to be turned so far that the 
point h (Fig. 1 1 , b) in the cylinder were twisted back to o, 
the tooth c would then fall past that point and lock the 
cylinder in that position. This is called overbanking, and 
to prevent this, a pin or stud is fixed in the balance 
cock in such a position as to be in, or very near, the 
centre of the vibration ; another pin, projecting on the 
opposite side from the rim of the balance which makes 
about three-quarters of a complete turn at each oscil- 
lation, prevents it from passing the projecting pin in the 


[Chap. IL. 

9ig, 11.— Horiiontal or Cylinder Escapement. 


cock. (The phrases commonly in use " balance makes a 
turn," " a-tum-and-a-half," &c., require perhaps a little 
explanation, as it is obviously impossible for a watch 
with any ordinary escapement to go if the balance 
swings round over a complete turn ; the meaning is that 
it makes a turn, &c., at each complete vibration, i.6., in 
its backward and forward arcs of motion added together.) 
These are called the banking stud and pin, and without 
them this overbanking would occur in the event of any 
sudden twist or jerk of the watch, or in the event of the 
balance vibrating too far, and so would stop the watch. 
Notwithstanding the general non-appreciation of the 
qualities of the horizontal escapement, the application of 
it to his watches, aud its superior performance as com- 
pared with the verge, gained for Graham a great reputation 
which soon excited the emulation of watchmakers here 
and on the Continent, and watches with this escapement 
were, a few years after its introduction, made in large 
numbers ; but the rapidity with which the cylinders 
were cut by the brass escape wheels, and the wearing of 
the brass pivot holes (which disturbed this escapement 
very quickly) prevented it from being universally used. 
These disturbing causes were, to some extent, overcome 
by the application, by Graham and some of his con- 
temporaries, of ruby cylinders and jewelled pivot holes 
(rendered possible, as we have seen, by the introduction 
about this time of the art of precious stone drilling and 
working, by Nicolas Facio, of Geneva), but the intro- 
dnction of the duplex escapement by Pierre le Roy, the 
celebrated French watchmaker, in 1727, bade fair to 
extinguish it altogether ; and it was not until the latter 
end of the eighteenth century, after exhaustive theories 
on the subject, published by Berthoud, Jodin, and other 
French watchmakers, that the capabilities of the hori- 
zontal escapement became generally understood. 

The peculiar adaptability of it for very flat watches 
was soon perceived by the Swiss, who, having altered or 
modified its proportions, substituting a light and small 


steel wheel for the brass one, and quick trains for the 
slow ones previously used, adopted it universally in their 
watches, and it remains one of the characteristics of a 
Swiss watch to the present day. 

. 86. To plan a horizontal escapement : From the centre 
A (Fig. 12) describe a circle, of which the arc B c fomiB a 

segment, and draw the radius A b. If the escape wheel is 
to have 15 teeth (the naiial onmber), mark off intervals 
of 24° from a b (2i° being ^^^°), and draw radii irova 
the centre a to these points. Mark off 10° to the ri^t 
of each of these radiL (When the circle determining 
the height of the inclines of the teeth is found, these 
radii wiU determine the positions respectively of the 
heeb and points of the teeth.) 

Draw the line ry, and snppotdng it is required to 


give 20° of lift on each side of the cylinder, take off 10' 
from this line, and draw e y, cutting the radius A B at Z. 
Then I b being = half the lift is = half the height of 
the impulse plane o£ the tooth. Describe a 6, and join 
B a, a Q, Q 6. From the centre o describe the circle o Q j 
this gives the outside diameter of the cylinder ; taking 
off 1" from AX on the circle B c will give the thickness. 

Not less than 18** from the radial should be allowed 
at the back of the tooth for freedom of the heel if square, 
or flat; or it may be cut away concentric with the 
cylinder ; and the height of the plane of the teeth must 
not exceed one-tenth of the radius of the wheel or the 
cylinder would be too thin- The diameter of the cylinder 
to that of the wheel should be about as 1 to 9, and its 
acting part 200** of the circle, inclusive of the rounded 
lips. The lips should be rounded, as shown in Fig. 11, 
which adds a little to the lifting arc. 

To form the impulse curve of the teeth : from the 
middle of the planes draw perpendiculars, as at kh'y 
from the centre A describe the circle h A, touching these 
lines ; and from the points of contact of the tajigents 
with this circle draw the curves of the teeth touching the 
points and heels of the planes. 

87. Parts of the Horizontal Escapement. — The 

cylinder should be pitched (as in the diagram) so 
that the middle of the height of the impulse plane of 
the tooth shall pass through its centre ; this will give a lift 
acting equally before and after the line of centres, and 
ensure the least possible amount of drop. As the 
balance in the return supplementary arcs attains a con- 
siderable velocity before the incline of the tooth comes 
into action with either lip of the cylinder, it is obvious 
that with an impulse plane of a very small angle, or a 
very shallow tooth, the cylinder will have travelled some 
little distance before the plane will come into contact with 
its edge, or, at all events, the impulse given on its first 
coming into contact will be very slight until it has 
travelled a certain portion of the incline, owing to the 


cylinder travelliDg as fast as, or faster than, the wheel 
on its first being released from rest, whereas with a plane 
of a very large angle the friction will be excessive, and 
the impulse will, in consequence, be very feeble (as may 
be observed in escapements with badly planted cylinders 
and very short teeth). The thing to be desired is, there- 
fore, that the relative sizes of the wheel and cylinder 
shall be such that they will allow of a tooth which will 
give the necessary number of degrees of impulse (with 
the plane in contact with the cylinder throughout its 
entire length) with the minimum of friction. 

The lift, or the angle moved through by the cylinder 
while it is receiving impulse either way, should be about 

The next point to be considered is the form of the 
edge of the tooth for giving the impulsa The three 
principal curves adopted for this purpose have been the 
following : — 

1. l^at which, by moving the wheel through equal 
spaces, will give equal arcs of impulse to the cylinder ; 

2. That by which the power of the impulse is ren- 
dered proportional to the increased resistance of the 
balance spring ; 

3. That approximating nearly to a straight line, 
which is thought to give the longest arc of vibration to 
the balance. 

Of these three forms, the first is the one that has 
been found to give the best results, and is that which is 
now universally adopted in the best cylinder watches. 
The second of these, having a ereater curvature near its 
point (which would no doubt^e the correct form of 
tooth were the increasing inertia of the balance, which 
overcomes the increasing resistance of the spring, left 
out of the calculations), commences to give impulse in a 
very sluggish manner, gradually increasing in velocity ; 
and as the latter portion of the curve near the heel ap- 
proximates very nearly to that of the circumference of 
the wheel, and encounters little or no resistance from 


the cylinder, owing to its accelerated motion, it traverses 
this portion with extreme rapidity, and falls with a very 
heavy drop on to the edge of the cylinder, which occa- 
sions a very rapid wear of that part, besides other 
irregularities incidental to excessive drop. 

The straight incline is a very good form for the 
tooth, but is inferior to the first curve, which gives a 
uniform impulse throughout its length; this curve is 
formed with the radius of the wheel. 

The proportions directly dependent upon one another 
(and by altering any of which we would alter the per- 
formance of the escapement) are :— The strength of the 
mainspring, the diameter of the escape wheel, the height 
of the incline of the tooth, the diameter of tiie cylinder, 
and the size and weight of the balance. 

In the best Geneva-made cylinder watches the escape 
wheel is made small and flat, with a proportionate 
cylinder, and a small but comparatively heavy balance, 
and this construction gives the best results of which 
this escapement is capable. 

Large escape wheels and cylinders should be avoided, 
as they require stronger mainsprings and lighter 
balances, and are more susceptible to disturbing in- 
fluences, such as thickening of the oil, &a 

In the case of the balance vibrating too fox, a weaker 
mainspring or a heavier balance should be applied, but 
if the balance should still bank, after either or both of 
these alterations have been made, the escape wheel 
teeth should be slightly reduced near the heel by a 
process called " topping " (ie., polishing or grinding a very 
little off them in the turns). This, by reducing the im- 
pulse and increasing the drop, vtHI have the desired effect. 

It is usual to allow of a little more drop on the 
inside than on the outside of the cylinder to ensure 
freedom for the heel of the tooth when inside it; for, 
were this precaution not taken, if the cylinder were 
planted a little too deep, and the tooth were the cor- 
rect size, the heel would rub. 


One property which the horizontal escapement pos- 
sesses, and which renders it peculiarly adaptable for 
going barrel watches (or watches without the fusee ad- 
justment of the mainspring), is, that it is not so much 
affected by any change in the motive power of the watch 
as any other escapement, the frictional rest of the tooth 
on the cylinder exercising a compensating power over 
the extent of the vibrations, so that any addition to the 
motive force is attended with additional friction on the 
cylinder, while the balance is performing the supplemen- 
tary arcs of vibration, and so retarding it and compen- 
sating for the additional force of the impulses. This 
isochronising power was what recommended it especially 
to the Swiss, who saw the possibility of suppressing the 
fusee, of which they had never been in favour, and which, 
in fact, they never thorougljly understood. 

88. The Cylinder. — Of the various methods practised 
for running in cylinders, the following, described by Mr. 
W. G. Schoof in voL xv. of the Ilorological Journal^ is, 
from its sin^plicity, among the best, as it avoids the ne- 
cessity for using intermediate gauges, which are in them- 
selves a source of error, and should always be avoided 
where it is possible to compare the piece of work itself 
under construction with the corresponding parts into 
which it is to fit. 

Both cocks having been removed — " I then," says Mr. 
Schoof, " take a small piece of brass, the shape of the en- 
graving (Fig. 13) and screw it in the place of 
the bottom cock, so that one of the holes — 
there are six in the illustration, though the 
number is of no consequence — comes in the 
same place as the bottom jewel hole would. 
Pig. 13. The slit of course receives the screw, and the 
object of its being a slit, instead of a hole, is 
to allow of its being moved, so that one of the six 
holes can be brought into a proper position. The escape 
wheel and centre wheel are left in their places, and 
having selected a suitable cylinder I put it in the 


frame, its arbor going through the hole in the piece of 
brass, or temporary bottom cock upon which the body 
of the cylinder rests ; then I allow a tooth of the escape 
wheel to pass into the cylinder, and turn the cylinder 
round until the plane of the wheel is in the slit of the 
cylinder. If the cylinder sufficiently clears the wheel it 
is right ; if it is too high, I turn away from the body of 
the cylinder at the bottom. Then, previous to turning 
down for the shotilder on which the balance is to rest, I 
mark with a safe-edge file the exact height of the 
shoulder, just sufficiently above the bar of the escape 
wheel on the fourth wheel pinion, which will be seen by 
resting the cylinder on the temporary cock. The height 
of the shoulder being thus marked off, when the cylinder 
is in the proper height for the escape wheel, I then pro- 
ceed to shorten the lower arbor of the cylinder to the same 
length as the proper bottom cock is thick, care being 
taken to allow for such depth, as the end-stone sometimes 
lies below the surface of its setting. 

" If now both cocks, having their endpieces removed, 
are screwed into their places again, the whole length may 
be taken with a pinion gauge or douziime tool and trans- 
ferred to the cylinder, by shortening the upper arbor, 
when the cylinder will be ready for pivoting. In pivot- 
ing a cylinder to a very flat watch, it is sometimes 
impossible to get one with a slit sufficiently narrow, and 
in such case it is advisable to let part of the cylinder go a 
little into the bottom cock, to avoid weakening the 
cylinder too much, and this the intelligent watchmaker 
will know how to allow for." 

The above applies only to the replacing of a broken 
cylinder, and would more properly come under the 
heading " jobbing," but as the manufacture of the hori- 
zontal escaj>ement has ceased entirely in this country, and 
is fast declining on the Continent, and even in Switzer- 
land, I shall not go into it any further than to explain 
the process to be observed in making a cylinder. 

89. Making a Qylinder. — A piece of good steel wire 


having been selected, a little larger than its ultimate dia- 
meter is to be, and a little longer than the height the 
cylinder will occupy in the frame, it must be drilled up the 
centre in a small lathe, with a meeting centre in the poi)pet 
head; the cylinder should be annealed before drilling. 
After drilling, the inside must be broached from both ends 
with a nearly straight broach, and, the height having been 
ascertained, polished till it will just admit a tooth of the 
escape wheel. The outside of the cylinder should be turned 
on an arbor and polished till it will just pass between two 
of the teeth. The height for the slots may be measured 
by placing a little red stuff on the edge of the teeth of 
the wheel, and touching an upright pointed peg 
previously forced into the lower jewel hole, with the teeth, 
or by means of the tool shown in the engraving (Fig. 14). 

Fig. 14.— Tool for measaring heights. 

Having cut the opening and slot, and rounded and 
polished the lips to the required shape, the hardening and 
tampering of the cylinder must be effected ; this may be 
done well in the following manner. Place the cylinder 
on end in the centre of a small copper vessel having a 
tightly-fitting cover and, having filled it (and the cylinder), 
with powdered charcoal, bind it tightly round with copper 
wire, filling the interstices with soap in order to ex- 
clude the air. Then heat to a dull red in a charcoal fire, 
and harden by dipping the whole into soft water; the 
cylinder may then be taken out and (after trying its 
temper) tempered by either boiling in oil or letting down 
to a blue colour ; the former is preferable as it ensures a 
more equable temper. The plugs should be made of the 
same steel as the cylinder of a smaller size and formed 
solid with the arbora. Having hardened and tempered 
them, they should be turned to almost fit the ends of tjie 
cylinder, when a few taps of the hammer will be sufficient 


to fix them in their places. The hammering of them in 
may be done either by holding the cylinder in a clamp 
shaped for that purpose, or with the stopping held in a 
clamp, and the punching administered through the inner 
edges of the cylinder with a punch of a suitable shape. 
The balance being colleted on (the collet being fixed by 
means of the above punch) and the pivots being finished, 
place the cylinder in the frame and try the drops of all 
the escape wheel teeth, and see that the cylinder is 
perfectly upright, and that sufficient freedom is left in 
the openings for the teeth and the wheeL 



90. Principle of the Escapement. — There exists a 
good deal of uncertainty as to who the inventor of this 
escapement really was. The idea, without doubt, origi- 
nated from the escapement of Dr. Robert Hooke, invented 
near the close of the 17th century. 

It was constructed in its present form by Pierre Le 
Roy, but his claim to its invention was disputed by 
Dutertre, another French watchmaker; it was intro- 
duced into this country by Thomas Tyrer, after whom it 
was also called Tyrer's escapement. 

The escapement as at first constructed had two escape 
wheels on the same staff or axis (whence its name), one 
of which was used for giving impulse, and the other, 
called the wheel of arrete, prevented it from running 
past the staff when the impulse tooth escaped from the 
pallet. The escape wheels were afterwards, as now, 
made in one, having the impulse teeth projecting from 
the plane, as shown in elevation in Fig. 15, I. 

Fig. 15 is a plan of the duplex escapement ; A is the 
escape wheel ; B, the ruby pallet fixed in a steel disc. 


and the ruby roller. The wheel is travelling in the 
direction of the arrow, and, the tooth h having escaped 
from the roller notch, the impiUse tooth c is giving 
impulse through the pallet B. When c escapes from 
B the tooth a will fall on the roller and rest until the 
return vibration brings the notch of the roller c round 
to its point, when it will fall into the notch, sustain a 
slight recoil, and, contributing a small amount of impulse, 
pass out as at ^ (^ig* 15> II-)> when the pallet B will 
have moved sufficiently round to allow the tooth/ to fall 
upon it and give the long impulse as before. The two 
impulses are technically called the great and small lift. 

The duplex escapement is not well adapted for going 
barrel watches from its extreme sensibility to variations 
in the motive force, and it was no doubt from this cause 
that Le Roy and the other French watchmakers quickly 
discarded it in favour of the cylinder escapement (which, 
as I have before remarked, possesses a regulating quality 
in itself), although they made many attempts to neutra- 
lise the effects of this irregularity by making the ruby 
roller extremely large, etc. 

As the whole secret of obtaining isochronous vibra- 
tions of the balance lies in its spring, the fewer obstruc- 
tions placed upon the free vibrations of the escapement, 
in the way of friction, etc., the better will be the per- 
formance of the watch ; and this is the secret of the good 
going of well made Duplex, Chronometer, and Lever 

Theoretically fche ruby roller cannot be made too 
small, but it is found in practice that it must be made 
of a certain size in order to obtain a safe rest, a suffi- 
ciently deep notch, and adequate strength to withstand 
breakage, and to prevent any bending of the staff, &c. ; 
it should, however, be made as small as possible, subject to 
these conditions being fulfilled. 

91. Angular measurement of Lift — ^The amount of 
the great lift and the length of the pallet from its centre 
are dependent upon the diameter of the impulse wheel. 


reUtively to ita distance between the centres. Of course, 
if th.e wbeel have less or more teeth, the pallet would be 
longer or shorter respectively. 

Fig. IS.— AngolftE Meai 

Fig. 16 is a diE^ram showing the angular measure- 
ment of the lift of pallets of respectively ^, I and ), the 
radius of the impulse wheel, a is the centre of the 


balance staff, and b the centre of the escape wheel The 
relative size of the wheel, it will be seen, affects the depth 
of the intersection of the tooth in the arc traversed by 
the point of the pallet. The desideratum is to ensure a 
safe intersection and a sufficient lift, with the impulse 
given as near as possible to the line of centres. This is 
best secured by giving a lift of from 30* to 35*, which 
would require a pallet of about a third of the length of the 
distance between the centres, with a roller of a little less 
than a third in diameter of the distance between the 
points of two resting teeth. (See Figs. 17, 18, and 19.) 

92. Banning of the Wheel — The train used in 
duplex watches is invariably the 18,000, as in the 
chronometer, and the balance usually vibrates nearly a 
turn. Overbanking cannot take place with this escape- 
ment as it does with the cylinder and the lever; the 
effect of the balance vibrating too far will cause the 
escapement to "run," i.6., two or more teeth will escape 
at one vibration, causing the watch to gain a few seconds, 
as is the case with the chronometer escape. 

Various methods were tried to prevent this running 
or tripping of the wheel. The old-fashioned plan was 
to fix a stud or pin on the balance staff just above the pallet, 
having a slot cut in it into which a pin, fixed in the 
staff, projected, allowing it to move a quarter of a turn. 
This stud had a sort of pallet projecting from it, and, if 
the balance moved more than half a turn either way, 
this pallet came in contact with a banking stud or pin 
fixed in the plate. 

Another ingenious plan was that of having a loop 
formed in the outer coil of the balance spring, which, if 
expanded too far by the extra extent of the vibration, 
came in contact with a pin in the arm of the balance, 
and thus controlled the vibration. But this loop in the 
spring was found to be very much in the way of its 
isochronous adjustment, since it prevented the length 
of the spring from being altered in the pinning in. 

93. Duplex Escapement — ^No modem duplex watches 


'^-ihS'^^-SS.^: . .„ 

Lmfftiiot pftllf 


have any contrivance to prevent running of the escape- 
ment; its regular performance being best ensured by 
careful wearing on the part of the person who carries 
it. It is on this account, and because of the delicacy of 
its parts, not weU adapted for persons of very active habits, 
or for those who hunt, ride, &a, but in the hands of 
a careful wearer it is capable of very good results, in fact, 
results such as few lever watches would giva A good 
deal has been said about the liability of the duplex 
escapement to set, but my experience is that a well-made 
duplex watch never stops from this cause in the course 
of ordinary wear. This setting of the escapement is 
occasioned by the balance being brought nearly to a stand- 
still by a jerk or twist given to the watch while in the 
position of the escapement in beat with the motive power 
off, and the long tooth falling on the roller while it has 
little or no motion, and so stopping the vibration. 

Thirty years ago, many of the watches made in 
London were exported to America, and, as the United 
States Government imposed a duty on completed watches 
but not on those in an incomplete state, most of the 
watches sent there had no cases, the movement being 
completed ready for putting into cases. Most of the 
movements were full plate ones and many of them had 
duplex escapements, the duplex escapement being then 
in great favour for what were called fine watches. 

When we consider the delicacy of this escapement, its 
unsuitability for a full plate watch, and the way many 
of these watches were made, we may easily understand 
why the duplex escapement got a bad name in that 
quarter of the world, and also how it was the Americans 
took to machinery, and made watches themselves. It 
must have taken a good deal of ingenuity to devise so 
thoroughly bad a watch as a full plate duplex, and what 
was bad in the original construction was soon made 
worse by the American jobbers and the fitters of these 
movements to the cases. The consequence has been 
that an escapement which is capable of and has given 


excellent results, has gradually gone out of favour, and 
almost out of use. 

94. Escape Wheel — ^The duplex escapement has, as 
may be seen, few parts ; the escape wheel, pallet, and ruby 
roller may be said to form the escapement. The escape 
wheel, having two sets of teeth, is difficult to make, and, 
as it is of the utmost importance that it should be perfect 
in all its parts, special tools and skill are required for its 
production, and chronometer and duplex wheel cutting 
has long been a distinct branch of watchmaking. This 
wheel should be made of the very best and hardest brass ; 
the long or resting teeth are sometimes cut with radial 
faces, and sometimes tapering to a point back and front 
from the rim of the wheel The point of the face of the 
impulse teeth should be exactly between two of the rest- 
ing teeth, and those teeth should stand upwards out of the 
plane of the wheel. They are in the form of a triangle, 
with the faces undercut at such an angle that when the 
points are at the commencement of the arc of intersec- 
tion with the pallet, they are parallel to its face ; if they 
are radial they will fall on the point of the pallet instead 
of its face and cut very rapidly. The wheel should be as 
large as possible, just freeing the arbor of the fourth 
wheel pinion. 

95. The Escaping Angle. — Watchmakers have dif- 
fered greatly as to the best escaping angle for this 
escapement, some advocating a shorter and others 
a longer lift. But it has been generally considered well 
in this, as in other escapements, to avoid extremes, 
and the lift prescribed, of from 30° to 35° at the 
balance, has been found to answer best, giving as 
long a free arc as possible consistent with avoiding the 
tendency to set. The proportions given in Fig. 19 are 
what are usually adopted in the best escapementa In 
Figs. 17 and 18 are shown wheels and pallets giving 24** 
and 31° of lift respectively, but with such proportions the 
intersection of the impulse teeth with the pallet is very 
small indeed when the freedom comes to be allowed for 


(which would reduce the acttud lift about 4°), which it is 
not ia the diagramB, Added to this, the risk of setting 


Fig. 20.— ActUiD of wc&pe wliMl, with rallar, ic, d oaneot ptoportioii*. 

is greatly increased, and the OBcapement would have to 
be constructed with very great accamcy to perform well 
with such intersections, and even with the proportions of 
witeel and pallet given in Fig. 19 the work mnst be very 



[Chap. XL 

I ' 

well dona The rollers in the three diagrams (Figs. 1 7, 1 8, 1 9) 
are about the size used in the best duplex escapement, 
and allow of a secondary impulse of 97" at the balance 

./' \ 




Fig. 21..— Action of wheel with too large a roller. 

or 6° at the wheel. This, taken off from the 24° through 
which the tooth moves at each alternate vibration of the 
balance, leaves an effective lift of 18** at the wheel, less 
about 2°, which it is necessary to allow for drop and 
freedom of the pallet in passing and re-passing the points 
of the two impulse teeth, whose diameter it intersects in 


the Bpace between tliem. If the centres are planted 
ferther apart there ia a very shallow intersection of the 
roller, which causes a good deal of eng^ing friction, 

-Aotlon ot wheel <rltli loo inull a rollsr, 

and is a Bource of danger from tripping, &c., and if there 
ia much play or shake in the pivots. Fig. 20 shows a 
wheel acting widi a roller of proper size. The impulse 
tooth t has jiist escaped the point of the pallet, and the 
resting toolJi r has just reached &e outside of the roller, 
against which it will remtun until the return of the notch, 


into which it will then drop, and, after sustaining a slight 
recoil, it will travei*se the aro a, a', contributing the 
small secondary impulse to the balance and escaping 
at a', by which time the impulse tooth t' will have 
advanced to t" and fall on the pallet, which will have 
arrived at that point, thus losing none of the impulse in 
drop, &a Fig. 21 shows a wheel acting with too large a 
roller, or rather it shows that it will not act at all, as the 
tooth ty having traversed the impulse arc as far as it 
can until its prosress is arrested bv the tooth r 
coming in contact ^th the roller, is held In that position 
until the return of the pallet^ which will then catch it 
and prevent any further action of the wheel. 

On the other hand, with a roller excessively small, as 
shown in Fig. 22, th^re is an excessive loss of impidse 
arising from the large amount of drop either on the 
pallet or roller, or on both ; and by moving the pallet 
round to decrease the drop on the pallet we increase that 
on the roller, and vice versd. 

Some writers on the duplex escapement advocate a 
large amount of drop in order to get the impulse close to 
the line of centres, and to avoid engaging friction ; but my 
experience is that all unnecessary drop is a direct evil, 
causing a more rapid wear of the parts from the jar 
consequent on the impact^ and a falling-off in the arc of 
vibration, and all that is necessary is to secure a safe 

96. Wheel Teeth. — In making this escapement 
the wheel is taken as the basis of construcfcion, from 
which all the other proportions are taken; if the 
radius of the impulse teeth be '7, that of the wheel, 
the lifting angle, will be about 35°, which experi- 
ence has proved to be the best for this escapement. 
The impulse teeth should be as wide as the arc described 
by the pallet will admit of, as should also the pallet, 
which must be flat on the face, like the pallet of a chro- 
nometer. The stone is often set in a shaped pallet, but it 
is better to set it in a roller, such as the roller in a lever 


escapement, as it is more easily made, and there is less risk 
of the stone being broken or loosened when the roller is 
turned on the staff to adjust the drop. The size of this 
roller is unimportant, but it should be large enough to 
afford a secure setting for the pallet stone. 

If the wheel teeth are quite true to each other, which 
can be tested by putting it on a true arbor, and touch- 
ing the tops of both sets of teeth with a peg, it may be 
taken for granted that the wheel is correctly cut. There 
is no other method of testing it but by its action when in 
the frama It may be crossed out with three arms, and 
these arms and the centre should be made as light as 
possible, the rim being bevelled off at the back for the 
same purpose, and it should be carefully poised. 

Though it has always been customary to polish the 
long teeth of this wheel, and the rim between them, there 
is more risk of spoiling the wheel by polishing the teeth, 
and less reason for doing so, than there is in polishing 
the teeth of the chronometer wheel. The teeth are polished 
on a thin sUp of hard wood held in the vice, with a thin 
tin polisher, charged with fine red stuff, the edge of the 
rim in the spaces being burnished as in the chronometer 

The depth for the fourth wheel and escape pinion can 
be got in the depthing tool, and the distance be marked 
on the bar under the dial. The depth for the wheel and 
roller is of more importance, and should be marked off 
with the utmost care. If a piece of steel turned true to half 
the size of the roller be put into the depthing tool with 
the wheel, and the tool be adjusted until the tops of the 
teeth just free it, this will be the proper distance of the 
wheel from the balance. The depth can now be marked, 
and a small hole drilled, when the escapement holes will 
be ready for jewelUng. 

The importance of properly steady-pinned cocks, 
pointed out at p. 196, is of still greater consequence with 
this escapement, since the least movement or alteration 
in their position would render it useless. 


97. Pallet Roller.— The roller for holding the pallet 
and the escape pinion can now be roughed out; the 
pinion must not be long, and should have a deep hollow- 
cut in its face to prevent it from drawing away the 
oil from the pivot hole. After driving a piece of brass 
or gold on, and forming a collet, and fitting the wheel 
tightly on this collet, the pinion may be faced and 
pivoted. Now cement the escape cock on to a piece of 
brass, and turn out in the mandrel a sink to free the 
impulse teeth, taking care not to turn it too thin. 

The steel for the balance staff should be of the best 
quality. After carefully hardening and tempering it, let 
die bfidance on to its place, and turn the seats for the 
spiing collet and pallet roller to size and polish them. The 
roller is made with a short pipe at the back ; it must be 
carefully fitted to the staff, as, although it must be tight, 
it should not be difficult to turn on the staff, the amount 
of the drop on the wheel tooth depending on the distance 
between the pallet and the notch in the roller. 

After pivoting the staff, turn down the lower part 
of it from the pallet roller to receive the ruby roller. 
This is a delicate operation, and requires great care, as 
the staff is at this part no thicker than a pivot. It is 
reduced on a bed in a rimmer in the turns with a steel 
polisher and sharp stuff. A small ferrule, a weak bow 
with a fine hair, lessen the risk of breaking. The roller 
must be fitted well, but easily, as if it is tight it will 
break, and if too loose it will not be quite true when it 
is shellaced on the staff. 

The diameter of the roller should be about '29 of 
the distance between the points of two of the resting 
teeth of the wheel, but the size may be varied a little to 
suit the pitch of the wheel. Escapement makers usually 
have a few rollers which have some defect, by which they 
get the exact size of the roller required ; there being no 
absolute size for the roller, the planting of the escapement 
to some extent determines it. 

The notch must not be too wide, as if so the recoil of 


the wheel is increased when the tooth falls into it in the 
returning vibration of the balance (a wide notch also 
increases the liability of the escapement to trip), but the 
teeth should have shake in every position of the notch to 
the wheel. The edges of the notch should be slightly 
rounded off and the roller well polished. 

The roller is fixed with sheUac, a brass cap a little 
larger than the roller being fitted to the end of the staff 
below it, and fixed in the same manner. This cap 
strengthens the roller and staff, and helps to secure the 
roller in its place. 

The balance must be rivetted before the roller is 
shellaced on, and the notch kept in such a position that 
the timing screws at the balance arms will be in a line 
with the pendant when the watch is in the case, as it is 
often necessary to alter these screws to bring the watch 
to time in positions, and with the shorter and more 
uniform vibrations of the balance with a duplex escape- 
ment, there are not the same objections to putting it a 
little out of poise that there are in a more detached 
escapement, where the balance has a longer arc of 

The heat for melting the shellac that fixes the roller 
and cap must be communicated to the staff by holding 
the arm of the balance in a pair of long-nosed pliers or a 
Swiss pin-holder, filed so as to grip the balance arm. 
The pallet should be close to the rim of the wheel, but 
free from it in all positions ; it should be in a radial 
line from the balance staff, and the pallet stone should 
be parallel to the acting faces of the impulse teeth of the 
wheel at the commencement of the arc of intersection. 

These impulse teeth should be accurately cut, and at 
equal distances from one another. The excellence of this 
escapement depends more on this than on anything else. 

The pallet should just be free of the top of the im- 
pulse tooth when it is brought round to receive the 
impulse, and be the same distance from the same tooth 
when the next long tooth is resting on the roller. 


The impulse tooth should only have perceptible 
drop on to the pallet when the balance is passed gently 
from, the wheel, but the pallet must free the tooth if a 
peg is put inside the rim of the balance, and a tooth is 
allowed to escape while the balance staff is pressed to the 
wheel. If the wheel is tried all the way round, and 
each impulse tooth falls on to the pallet with the peg 
outside the balance, and the pressure, therefore, is in the 
direction of the balance pivots, there will be no fear of 
the impulse tooth missing the pallet, as the balance at 
this point will be travelling so fast that the pallet will 
get some distance in front of the tooth before it receives 
the blow, and a very little drop on the pallet of a duplex 
escapement materially shortens the arc of vibration of 
the balance. 

The pivots of both the escape wheel and the balance 
must fit the holes very well, and have only perceptible 
side-shake, and the end-shake must also be very little. 

Duplex escapements are said to wear rapidly, but I 
have duplex watches that have been going for more than 
twenty-five years with scarcely noticeable wear in the 

That some of them have cut and worn rapidly may 
be easily understood from their construction. The inf- 
pulse teeth, even in full plate watches, were as narrow 
as possible, although there was here no stint of room, 
with a thin, steel pallet (appropriately called a hook), 
rounded on the face and end. The wheel — not always 
of the best — had too much drop on the pallet, and the 
wheel cutting the pallet, and the pallet the wheel, it 
soon grew too short, and a new pallet being difficult to 
make, the jobbing expedient of knocking the cocks nearer 
to catch each other, finished the escapemeniv 

Cliap.ZII.] 155 



98. Early Forms of Escapement. — The principle of 
the chronometer or detached escapement was invented 
by Pierre Le Roy, about the year 1747. It was subse- 
quently improved upon by Berthoud and by John Arnold, 
and was perfected by Eamshaw. 

The first chronometer escapements had the passing 
spring fixed to the roller, but Arnold and Berthoud 
transferred it to the detent. The detent worked on a 
pivoted arbor, having a spiral spring round it to bring 
it back into position after it was released by the small 
passing piece or paUet. Eamshaw improved upon this 
by doing away with the arbor and making the detent 
and spring in one piece, as shown in Figs. 23 & 24. The 
escapement of John Arnold is thus described in the 
specification of his patent, 1782 : — 

" The tooth of the balance wheel ▲ [in the drawing 
accompanying the specification] is an epicycloid [not 
a cycloid, as described by some writers^ that acts upon 
the pallet B, which in the part of action is a straight line 
from the outer edge of the pallet to the centre of the 
verge. The scape wheel rests on a single pin, whilst the 
balance is vibrating until it is unlocked, to add new im- 
pulse to the balance." 

The unlocking of this escapement took place inside 
the wheel, the acting curves of the teeth being raised 
from the plane of the wheel, as in the duplex, for this 
purpose. The escapements of Eamshaw had different 
shaped teeth, and, as in those now constructed, the un- 
locking took place outside the wheeL In modem chro- 
nometer escapements the teeth are also raised from the 


plane of the wheel, but this has nothing to do with the 
locking, and is only to give a greater wearing; surfooe to 
the teeth and a light wheel. 



99. Freseiit rorm.— Fig. 23 iB a diagram of the 
escapemeDt as at present constructed, c, the escape 
wheel, is arrested in its 
course by the stop J, 
which is a hard atone 
fixed in the detent ; the 
pallet /Tfixed in the small 
roller i is just commenc- 
ing to effect the release 
of the wheel by drawing 
the detent out from it in 
the return vibration of 
the balance. When the 
tooth e is thns released, 
the pallet d in the large 
roller will have arrived 
at a positioQ a very little 
in advance of the tooth 
r, which will fall upon 
it, and thus give impulse 
to the balance. In un- 
locking, the detent bends ^ ^ 

at a, which is a spring J 

formed in one piece with 
it. The sprin); m n is 
fixed to the detent at m 
hy a screw. This spring 
projects a little beyond 
the nose of the detent 
proper, and the pallet /, 
in the return vibration, 
lifts this spring and 
passes it without its 
having any perceptible 
retarding effect; it is 
made very weak and usually of gold. The detent is 
' sprung slightly on to a atop at s, which prevents it 
from going too deeply into the wheel. This stop, which 

; jtttw 



is not shown in the diagram, is a screw fixed in a 
separate arm of brass screwed on to the plate. It 
was formerly formed of a solid piece of metal, which was 
screwed on to the plate, but this was altered partly on 
account of the trouble of adjusting it to the required 
depth and partly because, being rigid, the repeated blows 
upon it of the detent produced a certain (or, rather, an 
uncertain) amount of adhesion very difficult to calculate 
upon. It will be seen that this escapement, like the 
duplex, allows the impulse to be given only at alternate 
vibrations of the balance. 

100. Lift. — As with the duplex, the requisite amount 
of lift governs the relation of the diameters of the escape 
wheel and of the arc traversed by the impulse pallet to 
one another. In marine chronometers, with their slow 
trains and heavy balances, the amount of lift found to 
be best is about 45**, while in pocket watches the maxi- 
mum allowed is 40°. Escapements allowing of lifts of 
respectively 45" and 36° for marine and pocket chro- 
nometers are what are usually made by our best chro- 
nometer makers. With a lift of 45° the radius of the 
pallet arc is about half that of the escape wheel (with a 
fifteen-toothed wheel). The complete arc of vibration is 
about 430°, or, as it is generally expressed, a turn and a 
quarter^ in both marine and pocket chronometers. To 
avoid the risk of setting, the 18,000 train is used in the 
latter instruments, while the former have the 14,400 
train, principally on account of the beats occurring at 
intervids of half a second rendering it more easy to 
observe accurately; the slow train also allows of a 
heavier balance being used, which is a gain in an instru- 
ment that always remains in the ^ame position, the 
momentum, being greater, giving a steadier rate. 

As the relative sizes of the roller and escape wheel 
are dependent upon the lift, so are the proportions of the 
detent, the radius of the arc described by the lifting 
pallet, and the locking depth of the detent stop in the 
wheel dependent upon one another. 


It is evidently advantageous to have the locking of 
escapements as nearly tangential as possible (other con- 
ditions apart), and this condition would be best obtained 
in the one under discussion by having the stop fixed in 
the detent at such a distance from the end that it would 
intercept the fourth tooth from the pallet (inclusive), as 
may be seen in the diagram ; the point of the fourth 
tooth being only a very little in advance of the vertical 
radius of the wheel, while that of the one now resting on 
the stop is considerably behind that line. 

101. Long and short detents. — And here is in- 
volved the whole question as to the relative merits of 
long and short detents, which is, after all, merely a me- 
chanical question embraced in the theory of the second 
order of lever, which the following diagram (Fig. 25) 
and proposition will expound :— 

Let p A B represent a detent reversed ; the point A 
is the stop which is required to be drawn out from the 
wheel after it has arrested it ; p is the power of the 
lifting pallet to effect this; w is the weight (or the 
resistance of the spring and the friction of the wheelon 
the stop) to be overcome, and P is the fulcrum of the 
lever (or the bending point of the spring). 

Let A p = a and b p = 6. 



^^ 4 ^ 3 


Fig. 25. 

Then the equation of equilibrium is 

P a = w 6, 
or Moment of p about p = Moment of w about p. 


If two forces act on a lever, they will balance when 
their moments about the axis are equal, and it is evident 
that (assuming P and w in the diagram to be in equili- 
brium) by moving the stop A nearer to the fulcrum P we 
should overcome the balance in favour of P, and vice 
versd. With a given wheel and roller, therefore, which 
determines the distance A b, it is clear that the nearer f 
is to the stop a the less will be the resistance offered to 
the imlocking. On the other hand (as to afford a safe 
locking, the stop must intersect the wheel a certain dis- 
tance), the nearer p approaches A the longer will the 
lifting arc of the lifting pallet require to be, which would 
increase its diameter, or, by deepening the intersection 
with an arc of the same radius, give it a more oblique 
action, and also keep it longer in contact, the one retard- 
ing the vibration and the other interfering with the 
detached properties of the escapement. An extreme in 
the direction of very short detents should, therefore, be 

An extremely long detent would have to be made 
considerably heavier to give it rigidity, and — ^in order 
that it should fly back instantly on being released from 
the lifting paDet — ^the spring correspondingly stronger, 
both offering greater resistance to the unlocking and to 
the good jperfoi-mance of the escapement, in which all 
the parts for obvious reasons should be made as light as 
possible. Moreover, the centre of percussion or of oscil- 
lation of the detent, where the stop should act, would be 
below the steel piece wherein the stop is fixed, and some 
other arrangement would have to be made for the bank- 
ing of the detent, which would otherwise not stop dead, 
but be in a continual state of vibration. 

The proportions of this and every escapement cannot 
be determined by any mathematical rule until the 
functions of each part have been carefully considered. 

The other proportions of the escapement being fixed, 
the total length of the detent must be determined by 
(1) the distance of the stop from the free end and the 


depth necessary to ensure a safe locking, and (2) by the 
lifting power relatively to the resistance offered by the 
spring of the detent and the friction of the wheel teeth 
on the stop. 

The impulse roller is always made of the diameter 
of the arc described by its pallet, having a piece cut 
out of it, as shown at D, in Fig. 23, to allow the es- 
cape teeth to pass. This not only facilitates the setting 
out of the escapement, but is a necessary precaution 
against the wheel teeth coming in violent contact with 
the detent (in the event of the pallet action failing and 
the wheel running), which they would distort and perhaps 
break, as may be seen when the staff of a chronometer is 
broken by accident. 

102. System of Manufacture. — Although the prin- 
ciple of the ship's chronometer has undergone very 
little change since the time of Eamshaw, the system 
of manufacture has been greatly altered during the 
last forty years. Many of the chronometers belonging 
to Government and kept at the Greenwich Observatory 
are by the early chronometer makers, and, as these 
men made the movements themselves, they also made 
the style of their work as distinctive as possible ; 
their chronometers were generally smaller than those 
of the present day, many being made to go thirty hours 
only. When movement making became a distinct trade 
the sizes were made uniform, and a small and a large 
"two- day" movement was adopted in addition to the 
"eight-day " then more frequently made. The smaller 
size has gradually disappeared, and the large " two-day *' 
is now the standard ship's chronometer. This uniformity 
of size simplified the manufacture by systematising it, 
and greatly reduced the price of chronometers, as for all 
practical purposes the sizes of the various parts are the 
same, although made by different movement makers, and 
are approximately as follow : — The dial is 4 in. in 
diameter, the barrel being as large as the frame and 
box will admit, from 1*5 in. to 1*9 in. in diameter and 



•75 in. in height ; the great wheel is about the same 
diameter as the barrel ; this baiTel will take a spring of 
sufficient strength and length to keep a chronometer 
going for fifty-six hours, with a balance of 1*25 in. 
in diameter and weighing -35 oz. ; these relative sizes of 
the barrel and balance are found from experience to 
answer best, and are generally adopted. The size of the 
escape wheel should be about '56 in., and a roller half 
the size of the wheel will give a lifting angle of 45*^ at 
the balance, this arc being greater than is adopted for 
pocket chronometers, there being no chance of a marine 
chronometer setting from external disturbance. 

It is not usual for escapement makers to make their 
own balances (it would not pay them to do so), but there 
is no difficulty in it, if they have the proper tools and 
wish to make experiments. A balance for a " two-day " 
chronometer will be supplied the proper size and propor- 
tions by any balance maker. 

The escape wheel should be made from a piece of the 
best brass, well hammered, and a hole the required size 
be drilled and broached (but not after the teeth have 
been cut) carefully ; it should then be turned on an arbor 
to the above diameter, true, and flat on the sides to about 
•065 in. in thickness, when it is ready to have the teeth cut 
on it. The teeth must diverge about 20^ from the radial 
at their faces, in order that the point only shall come in 
contact with the impulse pallet, and that the tooth shall 
have a tendency to draw the stone on which it locks 
towards the centre of the wheel. 

The roller is made of steel, in the same way as the 
brass disc for the wheel ; care should be taken in broach- 
ing the hole in the roller as in the wheel, and it should 
be left the proper size at once, as broaching afterwards 
is likely to put it out of truth The wheel disc and the 
roller must not be turned on the common Swiss arbor ; 
they are not true enough, nor of a proper shape ; the 
arbors must be tempered ones, turned true after they 
\ been tempered. 


The roller must be as much as it will lose in polish- 
ing larger than half the diameter of the wheel, and a 
little thicker than the wheel disc — about -08 in. When 
the wheel is cut, the depth can be marked off and the 
frame jewelled. 

Now put the fourth wheel and escape pinion in the 
depthing tool, and, after adjusting it for a proper depth, 
mark it off from the fourth wheel hole by a circular score 
on the underside of the top plate. This depth is not very 
important, as the finisher will have to run it over again, 
and he need not work to the fourth wheel hole ; but the 
wheel and roller depth requires great care in pitching, 
as after the holes are jewelled they cannot be altered. 
The hole in the potance should be turned out in the 
mandrel, the wheel and roller mounted on temporary 
arbors and put into the depthing tool, and the tool 
adjusted until the roller lies between two teeth without 
any shake, but free. The tool must be held perpendicular 
to the plate when marking off this depth, and where the 
circle intersects the one already marked on the plate will 
be the place for the escape wheel, when a small hole 
should be drilled, and the frame jewelled for the escape 
pinion and balance staff holes. The jeweller must take 
his uprights from the holes in the potance and top plate. 

Cement the escape wheel on a piece of flat brass, and 
turn out the centre right up to the teeth, leaving no 
ring, so that half the thickness of the teeth stands above 
the plane of the wheel ut the top (as shown in elevation 
in Fig. 24). This sink lightens the wheel, while it leaves 
the points, or acting parts of the teeth, wide enough. 
The wheel may now be crossed out. Escape wheels are 
generally crossed with three arms. It is a common 
practice to polish the teeth of the escape wheel, and a 
worse practice could not be adopted; it is not only 
useless but mischievous, as perfect truth of this wheel is 
indispensable to a good escapement. The backs of the 
teeth are polished with a tin polisher, the wheel having 
pieces of braes fitted on each side, and being fixed in a 



swing tool, and the bottom of the space between the 
teeth burnished. If the escape pinion be not quite true, 
it must be made so by filing the centres of the arbor, 
after which new centres should be turned, as filed centres 
will not keep true if worked upon. Turn the arbor to 
its proper size, and drive a brass collet on for the escape 
wheel ; this collet must be quite close to the end of the 
pinion arbor. The fourth pinion should be pivoted a 
good distance above the pillar plate to allow of its wheel 
freeing the centre wheel and the third wheel pinion. 

103. Marine Chronometer Escapement. — The me- 
thods used for obtaining the lengths of arbors or staffs 
that have holes with end-stones in small work, are not 
available in large chronometers, and, it being important 
that these heights should be obtained with certainty and 
exactness, various appliances are used for the purpose. 

Fig. 26 represents a simple tool I made to take these 
heights with. It consists of a very small brass tube, 
somewhat resembling a pencil-case, with a 
pivot fixed in a short plug at one end, and 
into the other fits a piece of brass, friction- 
tight, with a pivot at its extremity. These 
pivots are small enough to go into the balance 
staff holes, and by di*awing the centre out of 
the tube until it is a little longer than the 
distance between the holes you wish to work 
26— *^' ^^® pivots can be put into the holes, and 
Tool for the cock or plate pressed into its place ; this 
H^kte."*^ will give the exact distance between the end- 
stones and the length required. But practical 
escapement makers rarely use tools for this purpose. 

The pinion may now be faced and pivoted, and the 
arbor polished ; the wheel must be just free of the top 
plate, and very tightly fitted on the brass collet before it 
is rubbed on, otherwise it will not be safe. 

The sink in the top side of the wheel is finished grey 

with a paste formed by rubbing two pieces of blue stone 

together with a little oil on tke iace> oi ^w Vnqtc^ ^VjWcck \ 


or it may be greyed with a bow and drill stock and a 
piece of ivory fitted as a drill Before polishing the back 
of the wheel, a piece of brass should be fitted to the sink 
in front, and cemented into it to prevent the wheel from 
springing and being made too thin in the middle. The 
balance staff of a marine chrpnometer has the balance 
screwed on to it, instead of being riveted, and for that 
purpose a large brass collet is driven on to the roughed- 
out staff about half way. The staff should be treated in 
the same way as the pinion, and turned down for the 
impulse roller. As the balance must be kept close to 
the top of the plate, and the impulse roller just flush 
with the under side of the plate, the heights are easily 

The hole in the balance is usually larger than 
there is any need for the staff to be, and a seat is turned 
on the brass collet for the balance, which is fixed to the 
collet with two screws, whose heads are counter-sunk in 
the arms. The seats for the spring collet and impulse 
roller may now be finished, and the lower part of the 
staff, on which the small or unlocking roller has to be 
fitted, turned down ; this portion of the staff should be 
turned small, as the unlocking roller is not more than 
one-third the diameter of the impulse roller. The pivots 
should fit the holes, so that the loose jewel holes will 
fall off them with their own weight, and should 
be just long enough to project through them. 

104. Pivots.— -Conical pivots may be made 
much smaller than straight ones, as they are 
much stronger from their shape; but pivots of 
this form have a tendency to draw away 
the oil from the holes by capillary attraction, the 
pivot tapering gradually from the size of the pig. 27. 
arbor. To prevent this, in all good work, what is ^?"*^*i 
called a back-slope is made in the form of Fig. 27. 

This is generally looked upon as a mere ornament, but 
it is essential, as if too much oil is put to a hole the arbor 
will certainly draw it away if its tbick\v'i^^\s.\v'5X»^^^i»5»^ 


at the pivot end in this way. There has been much con- 
troversy amongst watchmakers as to the proper shape of 
pivots, but theorising on this subject without practical 
experience is worse than useless, and the first lesson ex- 
perience teaches is that the best pivots are those which 
stand best. 

A shape of pivot strongly recommended by several 
theorists and c scapement makers for reducing the side 
retardation is shown in Fig. 28. It looks very 
well on paper, but, as it is larger at the end, the 
actual side friction will be greater than with a 
conical pivot having its acting part sti-aight, and 
it is also weakest just at the bending or breaking 
Pig. 28. part, and if it is left flat on the end, as recom- 
mended, to equalise the side and end friction, 
the leasts contact with the edge of the hole will injure it 
sufficiently to prevent any good performance of the watch 
or chronometer. A good pivot should be so hard that it 
will break instead of bending, the portion of it that fits 
the hole being quite straight, and the cone should come 
close to the end of the pivot, as the acting part of the 
pivot is very short. The holes for conical pivots should 
be of the shape shown in Fig. 9, so that if they are not 
exactly opposite and parallel to each other, the edges of 
the holes will not, when in certain positions, grip the 
pivots and thus bind the staff". 

105. Locking Stone. — The impulse roller should now 
have the slot made in it to hold the stone : it must be cut 
in a radial line and the stone fitted tightly, the outer edge 
of the stone just coming to the edge of the roller. Many 
(J good chronometers have been and are made without a 

stone; they will go for years without the impulse face 
being marked if it is left hard, while if cut a little it is 
easily polished. Many of the old chronometers had no 
stone, and the pallet face was not in a radial line from the 
. , centre, but was undercut ; this was a mistake, as it gave 

I ? the wheel teeth too much drop, cutting the face of the 

pallet, and diminishing the arc of vibration of the balance. 


When the stone is let into the roller the crescent to admit 
the teeth of the wheel can be filed out ; there must be 
sufficient freedom behind the stone to allow the tooth to 
get away in case it misses the pallet, and the space in 
front of the stone should be only a little less than the 
distance between two teeth of the wheel. The small or 
unlocking roller has now to be made, and let on to the 
staff; its diameter should not be more than a third of 
that of the large one, as the unlocking stone must project 
beyond it, making its actual diameter nearly half that of 
the large one, and it must? not be more. This stone is 
sometimes made the same shape as the impulse stone, but, 
as it must lift the gold spring in passing and unlocking 
nearly the same distance on each side, it is best to be 
rounded off on each side like a thin tooth of a wheel. 

If a circle is struck off on the top plate the exact size 
of the wheel, and the wheel and balance staff, with the 
impulse roller on it, is put into the frame, there will be 
little space or shake between the two teeth of the wheel 
and the roller. Draw a line along the face of the third 
tooth which is to rest on the locking stone, and where 
this line cuts the circle describing the size of the escape 
wheel, will be the place for the locking stone. A straight 
line di'awn across the centre of the mark made for the 
locking stone from the balance will give the line on 
which the detent must be planted. 

1 06. The Detent. — The reasons given on pages 1 29, 1 30, 
point out the advantages of short detents, and they are 
fully borne out by the experience of modem chronometer 
makers. Detents are usually made '8 in. in length, measur- 
ing from the tip of the gold spring to the lower end of the 
detent spring proper, the length of the foot being of no 
consequence ; this length may be divided into four, one 
part for the horn, that is, from the tip of the gold spring to 
the locking stone, one for the spring, and two parts for the 
blade. The detent must be made of the very best steel, 
the hole for the locking stone drilled and broached to the 
size of the stone, and the hole in the foot for fixing it to 


the plate drilled and tapped. If the hole for the locking 
stone be brought over the mark made on the plate, and 
the hole in the foot on the line di'awn to fix its position, 
the hole in the plate may be drilled ; this hole should 
then be filed into an oblong slot, in order that the detent 
may be moved a little nearer to or farther from the 
balance holes, and the body of the screw for fixing it left 
plain and fitted to the slot. The stud into which the 
banking screw is fixed must now be filed up and planted 
along the outside of the detent, the place for the 
banking screw marked, and the screw fitted. The bank- 
ing stud must not be left too rigid ; the banking screw is 
put in from the side next to the escape wheel ; the pipe 
in the detent which holds the locking stone projects con- 
siderably beyond the blade of the detent (see Fig. 24), 
and rests, or banks, against the inside of the head of the 
screw ; the screw head should be large enough to just 
free the unlocking spring, and the pipe should rest on 
the outer edge of the head so that the banking may be as 
near the locking of the wheel looth as possible. As the 
screw must be very steady (but not hard to move) for 
the adjustment of the depth, it is best to make a slit 
with a fine saw from the end of the banking stud into 
the screw hole ; tlie sides of the stud can then be squeezed 
in a little and the screw will hold spring tight. 

Although there is here (unlike the escapement to 
pocket chronometers) plenty of room, the detent should 
be made as light as possible, the spring should be broad, 
and the blade stepped down from it, so that the wheel 
tooth, when resting on the locking stone, shall be in a line 
with the centre of the spring. This is the principal 
thing to be kept in view in filing up and planting the 
detent ; therefore the pipe must not be in the centre of 
the blade, but outside it, as the stone is for strength 
much wider than it is required to be for locking the 
wheel The detent spring must be weak enough not to 
offer too much resistance to unlocking the wheel, but it 
must be strong enough to reastV^endm^itomtke pressure 


of the wheel and also to return quickly to the banking 
screw when the unlocking stone is drawn out from the 
tooth. This power of returning does not depend alto- 
gether on the strength of the spring, for if the detent is 
long and heavy, it will not return quickly, even should 
the spring be a little strong. There should, therefore, be a 
correct proportion between the strength of the spring and 
the weight of the detent. If the detent when finished is 
held and deflected at the point half an inch, it should 
vibrate four seconds before coming to rest ; this, how- 
ever, is one of the things that experience alone can 



107. Preparing the Diflferent Parts. — In a three- 
quarter-plate pocket chronometer the diameter of the 
balance should in no case exceed that of the barrel, and 
the width of its rim should be '6 of that of the main- 
spring ; the length of the screw head and the thickness 
of the laminae, together, should be equal to the width of 
the rim. 

The escape wheel should be *45 of the diameter of 
balance, and the roller about '04 more than half the size 
of the wheel. The wheel should be made first in the 
following manner. Take a piece of the best brass, well 
hammered, fix it in the mandrel and centre and diill a 
small hole in it ; turn the brass flat on the sides, and cut 
out a disc very little larger than the wheel required ; it 
can then be put on an arbor and turned on the top and 
sides to the required diameter and thickness (about '04 
of an inch); the arbor should be perfectly true and 

When the wheel is cut, rough ou^^XifeTC^^-t^Va^srvss^ 


it a very little larger than it will be when finished, and 
somewhat thicker than the wheel, and drill the hole for 
the balance pivot half way between the centre hole 
in the plate and its inner edge, and a sufficient 
distance from the fusee to keep the balance quite 
free of the chain. If the fourth wheel hole coincides 
with the seconds pivot hole in the dial it may be worked 
from, but if not it must be stopped and another 
hole drilled exactly in the centre of the dial hole. 
The fourth wheel and escape pinion depth being got in 
the depthing tool, a score must be made marking this dis~ 
tance on the dial side of the bar, and the escape wheel and 
roller depth being obtained by fixing them on temporary 
arbors in the depthing tool (allowing the roller to lie 
between two of the wheel teeth without shake), this dis- 
tance should also be marked on the bar ; tlie point where 
the scores intersect one another will be the position for 
the escape hole. 

The cocks having been steady-pinned (page 196), the 
escape cock should have its thickness reduced in the man- 
drel to about '05 of an inch, and a circle sufficiently large 
to free the wheel and about two-thirds of its thickness 
should be turned away from the under side of the cock, 
and the cock must be stoned-up flat, when it will be fit 
for jewelling. As there is no boss for a regulator or index 
left on the balance cock, all the further preparation that 
is necessary is to turn it down till its top side is in the 
plane of the pillar plate ; that is, if the pillars are high 
enough to allow sufficient space for the cylindrical spring 
usually applied to pocket chronometers, if not, the cock 
must be left sufficiently high for that purposa 

The sink for the fourth wheel j^h^uld be turned out in 
, the bar before sending it to the jeweller to about -03 of 
an inch for his setting, otherwise he may make his setting 
too thick ; the end-stones to the escape wheel hole in the 
bar should be set in steel, the setting and screw heads 
being left projecting above the bar. The escape wheel 
should be cemented on to a piece of brass, and the upper 


side of it turned away in the mandrel to about half its 
thickness right up to the teeth, as directed for the marine 
chronometer escape wheel, to lighten it. The rim should 
be divided into three, and the wheel crossed out ; by 
clamping it between two small discs of brass in the swing 
tool the backs or hollowed parts of the teeth may then be 
polished with a thin tin polisher and fine red stuff, and 
the spaces between them burnished with a narrow bur- 
nisher. The escape pinion and balance staff can now be 
proceeded with: the pinion should first be shortened to a bout 
twice the thickness of the fourth wheel, which gears into 
it, and have a deep hollow cut in its face, to prevent the 
oil from the pivot hole from getting into it. The other 
side of the arbor must now be turned down and the 
collet for the wheel driven on to it ; this collet is usually 
made of soft steel, but gold is better. 

There is little difficulty in getting the height, as the 
pivot must be as close to the pinion as possible, and the 
wheel let on to the collet so that it will come close to the 
pillar plate. When the lower pivot is the proper length, 
the total length for the pinion can be obtained by mea- 
suring from the outside of the jewel holes with the end- 
stones taken off with a douzihme tool, which, for greater 
accuracy, should have a vernier slide gauge attached to 
it. Wlien the balance staff is roughed-out and tempered, 
the balance can be let on to its seat, and the staff 
shortened until the balance is sufficiently free of the 
escape cock to allow for the length of pivot that will go 
through the hole. When the pivot is made, the arbor 
may be polished and the roller opened until it will go on 
to its place on the staff ; the staff must be turned down 
small below the roller for the small unlocking roller, 
which, as in the marine escapement, is about a third of 
the diameter of the large one, when this roller can also be 
made and let on to the staff. The impulse roller can be 
cut with a file or cutter in a radial line, and the stone 
fitted. It must fit tightly and be perfectly square on its 
face to the balance staff, a crescent being cut out of the 


roller's circumference to permit the tooth to fall on the 
pallet ; two-thirds of this crescent should be in front of 
the pallet. The roller, if the proper size, can now be 
hardened and tempered (it should be let down to a blue), 
and finished, the stone being fixed in with shellac. 

The point or edge of the pallet must be exactly in the 
circumference of the roller ; if a hole is opened in a piece 
of brass to the size of the roller the pallet may be pressed 
against the side of the hole while the shellac is warm, 
which will bring the outer edge of the stone in this 

The small roller may be proceeded with in the same 
way : the notch for the stone must be cut nearly to the 
centre ; the stone should be of the form of a very thin 
wheel tooth, and should not project to more than half the 
diameter of the impulse roller. Before this roller is 
hardened, a part of its circumference is filed flat, to avoid 
the risk of the pliers slipping on to the stone while the 
roller is being shifted to regulate the drop of the wheel on 
the pallet. 

108. Spring Detent. — The spring detent is, of course, 
the important and critical part of the chronometer escape- 
ment, and it would be a mere waste of time for any but a 
good workman to undertake the making of one. Draw a 
circle the exact size of the escape wheel on the plate, and 
put the wheel and the balance staff, with the roller on it, 
into their places, by holding the wheel with the fingers in 
such a position that the roller is free of the tooth on each 
side of it ; a line may then be drawn on the plate with a small 
point along the face of the tooth that would be resting on 
the locking stone of the detent ; the point where this line 
intersects the circle drawn on the plate shows where the 
locking stone must be placed, and a straight line drawn from 
the centre of the balance hole across this point will show the 
exact position in which the detent must be planted. The 
best cast steel should be used for the detent, as, if it 
will not harden at a low temperature, it will be useless. 
I have found Huntsman's steel the best for this and other 


purposes where a similar quality of steel was required. 
The length of this detent is easily determined : it must 
not extend beyond the edge of the plate, and must be 
made short in order to secure lightness. If the foot of 
the detent be stopped down until the lower part of the 
steel touches the fourth wheel, the filing-up can be pro- 
ceeded with. When the greater part of the steel is 
removed from what is to form the blade and spring, the 
hole in the foqt can be marked and drilled, and the hole 
for the screw in the plate also drilled ; the distance from 
this hole in the plate and the mark already made to deter- 
mine the place for the locking stone can be marked 
off on the detent and the hole for the stone drilled. 

The blade of the detent must be kept free of the 
underside of the escape wheel, but the spring should be 
twice the width of the blade, half the width standing 
above the blade in the plane of the escape wheel ; at the 
place where the hole is drilled for the locking stone a pipe 
is left, so long that it is only free of the fourth wheel ; this 
pipe rests against the banking, and also enables the locking 
stone to be securely fixed. It must be outside the blade 
of the detent, for the same reasons given for that of the 
marine escapement. 

109. Basking Piece. — When the detent is so far 
advanced, the banking piece should be made. The usual 
form now in use is what is termed the horse-shoe, which 
may be made in the following manner : — 

Turn out a sink in the pillar plate round the balance 
staff about -03 of an inch in depth, and as large as it can 
be made without going into the sink for the centre 
wheel. Take a piece of sheet steel, thicker than the sink, 
as the projecting arm that is to receive the banking screw 
must go below it ; make a hole in the steel piece large 
enough to free the unlocking roller, and a circle on it the 
size of the sink, then file and turn it to fit the sink, 
leaving an arm projecting for the banking screw. This 
arm should be placed in a line with the detent and close 
to it, and the banking piece should be fixed to the plate 


with two sunk-headed screws, and both screws and 
banking piece be somewhat below the plate when finished. 
Make the arm of the banking piece just free of the 
fourth wheel, and the place for the screw will be seen 
from the mark already on the plate for the locking stone. 
The arm must now be drilled and tapped, and fitted with 
a small gold screw, having its end turned ofi" to free the 
gold unlocking spring ; this screw must be fitted very 
tight, and the banking piece should be very well fitted 
to prevent any possibility of its being shifted, which 
would alter the depth of the locking stone in the wheel 
The hole in the foot of the detent is filed into an oblong 
slot, and the screw to fit it to the plate is made with 
a plain part to fit exactly the width of this oblong. If 
the locking stone be temporarily fixed in the pipe with a 
half-round pin, and the escape wheel and balauce, with 
only the impulse roller on the staff", and the banking stud 
be put into their places, the position for fixing the detent 
in can be tried by moving it towards the roller until 
the escape wheel tooth is resting on the locking stone, 
and the roller is just free of the tooth on each side 
of it. 

The position of the banking screw had better be tried 
before the stud is hardened, as it will be easier to alter 
it then than afterwards. When its position is deter- 
mined, and the detent fixed by the screw in its foot, 
the hole for the steady pin can be drilled, and after 
drilling and tapping the small hole in the blade for the 
screw which fixes the unlocking spring, it can be 
hardened; it is best not to make the spring too weak 
before hardening, but to reduce it afterwards. The best 
way to harden it is to tie two pieces of thin brass, one on 
each side of the spring, with fine binding wire, using a 
charcoal fire and water ; the tempering may be done by 
blueing, but the safest way is blazing in oil, letting the 
oil ignite freely, and after the flame is extinguished, 
allowing the detent to cool in the oil, when it may be 
finished. It should be made flat and square with a steel 


polisher and oilstone dust, and reduced to the necessaiy 
dimensions (of course, the lighter the blade is made, the 
thinner the spring may be), after which it may be 
finished, — the grey parts with fine emery and brass or 
tin polishers, and the polished parts with zinc polisheis 
and diamantine. 

After having ascertained the angle at which to set the 
face of the locking stone (it should be set back so that 
the tooth of the wheel will draw it up to the banking 
screw head), it may be permanently fixed in the pipe 
with a half-round pin and shellac ; it should not be quite 
upright, but inclined at such an angle that the lower part 
of the wheel teeth lock on it close to the detent. This 
will prevent any springing of the detent, by getting the 
pressure near the middle of the spring. 

110. The Unlocking Spring. — The unlocking spring 
is usually made of gold, hammered very hard ; it is at- 
tached to the detent by a very small screw close to the 
spring ; and, instead of a hole in the end for this screw, a 
notch is cut, the size of the screw, to allow of the spring 
being adjusted for the unlocking lift. The gold spring 
must not be wider than the blade of the detent, except at 
the ends, where it is also left thicker than in the middle, 
and it must be weaker than the detent spring. It should 
lie close alongside the detent, but should only touch it 
where the end is screwed on, and at the extremity of the 
horn against which it rests ; it is slightly bent, or sprung 
on, so as to return to its place against the horn when the 
unlocking pallet moves it away in the return vibration 
of the balance ; its end must be carefully filed up and 
finished, and it must lie horizontally in order that the 
unlocking stone may come fairly on its side. The lift of 
this spring must be only perceptibly less in passing than 
in unlocking, since, if the passing lift is very little, the 
unlocking will not be ceiiiain, and the spring should be 
only left long enough to carry the detent just perceptibly 
beyond the point where the wheel teeth drop from it. 

When the escapement is completed the detent is sot 


on to the banking ; this may be done in the process of 
polishing, or, if this is thought dangerous, by. heating 
a piece of steel or brass wire in the flame of a spirit 
lamp, holding one end in the vice, and bending the 
spring of the detent against a part of it that is not too 
hot, and pressing it there until it becomes a straw colour, 
allowing it to cool in that position, when it will retain 
the bend or set thus given to it; there is no risk of 
breaking it in this way. 

It must be remembered in making the chronometer 
escapement that, the balance being more detached from 
the train than in any other escapement, and the inter- 
section of the parts that give and receive the impulse 
being consequently less, the greatest accuracy must 
be observed in making and trying all the actions and 



111. Double Roller Escapement — Fig. 29 shows the 
plan and elevation of the lever escapement, a is the 
balance, B b' the lever ; c, the escape wheel ; ni, 7n\ the 
pallet ; t, the ruby impulse pin ; c, the impulse roller ; c, 
the small roller ; v, the guard ; «, the escape wheel pinion, 
its staff being cut away in the diagram (Fig. 29, ii.) 
to show «', the pallet staff. Pai-t of the balance rim is 
broken off to show the parts underneath better. 

This figure is a diagram of a double roller escapement, 
the action of which is the same as that of the single or 
" table " roller referred to further on. 

The balance receives impulse at each swing or half 
vibration. The escape wheel tooth having escaped the 
inclined plane at m', the tooth on the opposite side at m 
has fallen on the locking face of the pallet, where it will 


remain until the balance returns in its vibration, and 

. ritui. n. EEeratian 

the impulse pin (() coming into the notch will then move 
the lever (bb') over totthe other aide ; the escape wheel 
tooth in this motion, as the pallet at m is drawn out 


Erom the wheel, falling over the corner and sliding along 
the inclined plane, giving impulse to the balance through 
the lever notch and the impulse pin (t). The pallet is 
prevented from going too deeply into the wheel by a 
couple of banking pins (r r') planted in the pillar plate of 
the movement^ which are not shown in the elevation. 

The locking faces of the pallet are cut back at an 
angle from the radius of the escape wheel, so that the 
teeth may hold the lever in the position of rest against 
the banking pins until the return vibration of the 
balance releases it. A slight supplementary run, as it is 
called, of the locking face, past the tooth's point takes 
place after a tooth drops on to either face, occasioned by 
the tooth drawiiig the pallet still deeper into the wheel 
than when it at first di*ops on to it. 

As the lever passes the large roller (c), the guard (v) 
passes through a space cut out of the small roller (c'), 
and this guard is so placed that when the lever banks, it 
is just free of that roller. This is to prevent the lever 
from being thrown out of position to receive the impulse 
pin in its return, by any outward jerk or shake drawing 
the pallet away from the tooth which is holding it, the 
horns of the lever (i i'), also are prolonged, as shown, to 
guard against any possibility of the pin missing the 
notch. Any such jerk would cause the guard to strike 
against the edge of the roller, whence it would rebound 
into the position of banking. But this guard is useful 
only in an emergency, as, if either locking face were so 
insufficiently back-sloped as to allow of this occurring fre- 
quently, it would speedily affect the rate of the watch. 
The end of the lever (b) is only a counterpoise. 

If the importance of the lever escapement were to be 
measured by its all but universal application to English 
watches at the present time, its treatment should occupy 
a very considerable portion of a book of this kind. 
Enough has been written about it to fill a very large 
volume without exhausting the subject, but space will 
not allow of my doing more than barely summarising it. 

Chap. XIV.l 



Although the detaxjhed lever was invented as early as 
the year 1750, it was the last escapement that came into 
use in high- class pocket watches ; and as I am treating 
of escapements to some extent chronologically, I shall 
here discuss it, — the last of the watch escapements. 

112. The Back Lever. — The first escapement re- 
corded on the principle of giving greater angular impulse 
or lift to the balance through an intermediate piece 
or lever, was invented by Huygens, or TAbb^ Haute- 
f euille, the latter of whom 
published a description 
of it in the year 1722. 
This was identical with 
the Rack Lever (Fig. 
30), patented by Lither- 
land of Liverpool, and 
which I have before no- 

It consisted of anchor 
pallets similar to those of a 
recoil clock escapement, on 
whose axis was fixed a rack 
or segment of a toothed 
wheel which geared into a 
pinion on the axis of the 
balance, the balance being 
at no part of its vibration 
free, or disconnected from 
the train. This, apart 
from the friction of the 
rack and pinion, rendered 


Fig. 30.— Back Lever. 

good timekeeping impossi- 
ble. As shown in Fig. 30, 
the end of the lever terminates in a rack acting in a 
pinion on the axis of the balance. This rack cannot 
leave the pinion ; the only effect of an external shake 
causing the balance to vibrate more than its normal arc, 
would be to drive the pallet flukes deeper into the wheel 


or to the bottom of the tooth. There is no necessity for 
any safety action, no banking pins, and (as the pallet rest 
is circular) there is very little friction or resistiuice to the 

113. The Detached Lever. — About the year 1750, 
Mudge invented the detached lever escapement ; and it 
is to some extent a reflection either on the mechanical 
genius or the goodnature of the watchmakers contempo- 
rary with and immediately following him, that it was 
allowed to lie dormant for more than half a century ; 
although it is, in some of its parts, the model of the best 
form of the lever escapement. The spring detent escape^ 
ment of Arnold, and the cylinder, or horizontal escape- 
ment, invented by George Graham, were at this time 
attracting all the attention of watchmakers, and Mudge 
himself was absorbed in trying to obtain the aws^ 
oflfered by Parliament for the improvement of marine 
timekeepers, so that his lever escapement was neglected. 

In some of its proportions it was faulty and could not 
have given very satisfactory results, and he probably 
did not feel enthusiastic about it, as he seldom men- 
tions it in his correspondence, &a, and only made two or 
three such escapements. It was the same in form as the 
present double roller, the lever and roller action not 
having since been improved upon. Mudge's son gives a 
drawing of this escapement in a book published in 1799, 
in which the pallets (embracing five teeth of an escape 
wheel with ' twenty) resemble very much those of 
Graham's dead-beat clock escapement. They had no 
" draw," as it is called, on the locking faces (the rest being, 
as in the clock pallets, circular), and therefore any jolt 
given to the watch would throw the guard against the 
roller, and thereby interfere with and retard the motion 
of the balance. The defect in the locking may have sug- 
gested to Mudge the small roller, which offers less resis- 
tance to the balance by the safety pin coming in contact 
with it. 

However, when the above fault was corrected by the 


lockings being back-sloped, as in escapements made sub- 
sequently, the value of the double roller was soon mani- 
fested, and there is no form of the lever escapement so 
trustworthy, or that gives such good results ; but, being 
somewhat more delicate, it is more difficult to make, and 
consequently more expensive than its now only rival, the 
table roller. 

114. The Crank Lever. — Arnold's spring detent could 
only be applied to large watches, and, being extremely 
delicate and difficult to make, was very expensive ; while 

Pig. 31.-— Crank Lever. 

the cylinder escapement (much esteemed at first) soon 
began to lose favour in consequence of the brass escape- 
wheels then used cutting the cylinders very rapidly. 
When, therefore, Litherland of Liverpool resuscitated the 
lever escapement by taking out a patent for the rack 
lever, which had been improved by Julien Le Roy, but 
not hitherto made in England in this form, it became very 
generally made in this country, and made rapid headway 
against the other escapements. It had some good points, 
and, being simple and strong, was very well suited for 
the watches of commerce then made. But this escape- 
ment, although an improvement, did not long satisfy the 
demand for a better mode of transmitting the impulse 
to the balance, and Massey, also of Liverpool, shortly 
afterwards invented the crank lever (Fig. 31). In 
this escapement the roller is small, having a tooth 


• I 




exactly like the leaf of a pinion projecting beyond 
its circumference. This tooth acts in a square notch 
cut in the end of the lever, and the lever is formed 
like a fork, the two points of which act as safety-pins 
against the edge of the roller to prevent the lever from 
getting out of action with the tooth of the roller. 

This lever and roller action was not so good as that 
of Mudge, nor would the performance of the escapement 
have been so good had not Massey altered the <B^pe of 
the resting £aoes of his pallets, making them straight and 
giving them a certain amount of " draw." This ** draw '* 
of the pallet was all that Mudge's escapement required 
to make it what it now is, nearly perfect as an escape- 
ment. But although improved by having a ruby pin 
substituted for the pinion leaf, and the roller raised 
from its plane, or elongated, to allow of the guard points 
acting as before, the crank lever required to be made 
very carefully or, if not, the pin would sometimes catch 
on the points of the lever and stop the watch. The notch 
also had to be made wider in reference to the pin for the 
same reason, which resulted in too much drop in the 
impulse — ^another source of irregularity which a watch- 
maker will fully appreciate. These difficulties brought 
about various modifications until the introduction, by 
another Liverpool man, of what is now called the table 

115. Table Boiler. — This escapement, from its sim- 
plicity and general excellence, is what is employed in nine- 
tenths of the lever watches of the present day ; the only 
difference between it and the crank lever is in the roller 
action. The impulse pin, instead of projecting beyond 
the edge of the roller, is set within its circumference, 
standing up out of its plane. The safety action is effected 
by a very small pin, placed quite close to the bottom of 
the notch in the lever, and passing the roller in the same 
way as the guard of the double roller escapement, through 
a small segment cut out of its circumference in front of 
the impulse pin. 



116, Two-pin Escapement. — Another form of lever, 
invented by George Savage, a man who did much to im- 
prove the lever escapement 
and make its capabilities 
better known, is CEilled tb« 
" two-pin " escapement, al- 
though, I think, " three-pin " 
would have been a more 
appro]triate name, for the 
two pins referred to do only 
the least important part of 
the work, i.e., the unlocking 
of the wheel. (Fig. 32.) 

Savage saw there was a 
loss of power consequent cwi 
the double duty of the im- 
pulse pin of an ordinary 
lever escapement (which, be- 
fore receiving impulse, has 
to unlock the wheel through 
the opposite side of the lever 
notch and necessitates a 
certain amount of shake), 
and also on the unlocking 
action taking place before 
the line of centres of the 
lever and roller. He there- 
fore separated these func- 
tions, reversing the ordinary 
method of placing the im- 
pulse pin in the roller, and 
the notch in the lever, by 
cutting a very small notch 
in the roller and placing a 
small pin in the lever, occu- 
pying the place of the ordinary guard pin, for which 
purpose it also served. He then placed two pins in the 
roller at each side of the notch to effect tiie unlockiDg : 

Fig. 32.— Two-pin Lsrer E«ap»- 


and as the distance between these pins is equal to the 
chord of the lifting arc at that radius of the roller, the 
unlocking takes place at the line of centres ; and, with 
the proportions adopted in the figure (viz., 8^" on the 
planes, and IJ** on the locking faces of the pallet) the 
impulse \\'ill be given at J^ from the line of centres on 
the lever. But this escapement must be constructed with 
the greatest nicety, and is very difficult to make. It has 
also been found that the small pins, made generally of 
gold, cut the steel on which they act very fast. 

Various attempts have been made to prevent the 
wear of the parts by substituting a broad thin stone set 
in the roller for the two gold pins, and a small ruby pin 
for the gold one in the lever ; but notwithstanding this, 
and the fact that this is theoretically the most per- 
fect form of lever escapement, there are few of them 

117. The Double Boiler, as its name indicates, has 
two rollers on the balance staff, the large one carrying 
the impulse pin, as in the table roller, whilst the small one 
is used for a safety roller only. In an ordinary escape- 
ment, with a lifting angle of 30^ at the roller, the 
intersection is only just safe when the escapement is a 
good one and all the parts well made and jewelled; but 
if the pallet-staff have brass holes and less skilful workman- 
ship generally, pallets of higher angles and a longer es- 
caping arc are necessary. There are no proportions of the 
lever escapement upon which greater diversity of opinion 
exists than on the proper lifting angle of the pallets. 
Lifting angles of 15** may be found in old watches, while 
some modem makers advocate as low an angle as 6°. 
Now, as the driving planes increase in length with the 
lifting angle, and also become more divergent from the 
course in which the wheel is travelling, the friction 
increases, and in an increasing ratio as the planes ap- 
proach more nearly to lines of the wheel radii 

On the other hand a pallet of as low an angle as 6® 
would require a greater pro\iOTt\oTi^ ^\s\q>k£^ ^1 d<a^th 

I , 


on the locking faces, as the amount of locking, to be safe, 
could not be reduced beyond a certain point ; therefore 
the ejQTective part, or the angle of real lift, would be pro- 
portionately less, and it would be veiy difficult to make 
an escapement that would be safe with a pallet of so 
low an angle. Experience has decided the matter, by 
showing that it is better to avoid extremes and in favour 
of pallets of from 8° to 12^ of lift. 

As, in the double roller escapement, the guard of the 
lever approaches nearer to the axis of the balance, and 
(as has been demonstrated in the previous remarks on 
escapements) the nearer that guard, or pin, approaches 
the centre, the deeper will be the intersection and there- 
fore the greater the safety ; pallets of as low an angle as 
8° may be used. This is analogous to the relative sizes 
of the pallets and wheels of the duplex and chronometer 
escapements, and will be further treated of in my direc- 
tions for planting the guard pin. Practical watch- 
makers doubt whether there is sufficient advantage in 
the principle, to compensate for the actual drawbacks in 
the shape of the difficulty of arriving at the extreme 
exactness requisite in the making, and for the risks of 
accident or wear attending any of the different parts. 
It is not advisable to adopt such proportions for watches 
of an ordinary character, and many escapement makers 
are of opinion that pallets of 12^, with a lift of from 
32° to 35°, will answer best in all but the very highest 
class of watches. I believe, however, that in such a 
watch a double roller escapement, with an 8° pallet and 
a lift of 25°, will be the nearest thing to a perfect escape- 
ment to a pocket watch, as in consequence of the deep 
intersection of the guard with the small roller, a safe 
action is insured, which permits of a shorter locking, 
and, as the action of the ruby pin in the lever notch 
takes place, both in the unlocking and the impulse, 
nearer to the line of centres, there is no danger of the 
balance setting. In fact, it cannot do so if of a proper 
size. Thick oil on the pallet plaives. ^\VV Viia?^^ Vijss. '^^'^^ 


tlian with those of a higher angle, aad the balance will 
of course Lave a longer free arc. 

The Swiss make double roller escapements to all their 
high-claas watches, and indeed to a great many that 
cannot be included in that category that would be a good 
deal better with something less pretentious. 

118. Cole'a Beeilient Escapement — Another form 
of the lever escapement (known as Cole's Resilient) was 
invented by the late Mr. J. F. Cole, about twenty-five 

years ago. It was intended to obviate tlie evils arising 
from overbanking. (Fig. 33.) 

The escape wheel resembles an ordinary escape 
wheel acting against the faces of the pallet with 
the backs of its t«eth, the points being bent towards 
the locking faces. Some of the wheels have been 
actually made in this manner on account of the 
difficulty in cutting them. No banking pins are 
required ; after the tooth drops on to the locking face, 
the suppJementery run draws the plane into the wheel 
as far as the bend in the tooth, which acta as the 


banking. In the event of overbanking, the effect of the 
impulse pin striking the lever would be simply to drive 
the pallet plane still further into the wheel, when it 
would slide down the reverse slope of the tooth and 
cause the wheel to recoil, which (after the force of the 
jerk was expended) would then force the pallet up 
again into its place, thus preventing any damage to the 
escapement, such as the ruby pin or the balance staff 
pivots being broken, or, at all events, preventing the 
acceleration in the time of the watch which would be 
occasioned by the pin striking the outside of the lever 
on alternate sides and rebounding several times, which 
it would do were the bankings rigid. 

This escapement was much thought of when first 
introduced, but the danger of overbanking was ex- 
aggerated, not one watch in a thousand being injured 
from this cause; and as the escapement was a little 
more expensive than the ordinary one, on account of 
the difficulty of the cutting of the escape wheel, its 
manufacture was abandoned. 

There have been numerous inventions to deal with 
this fault of overbanking, mostly in the direction of 
spring bankings, a spring lever, &c., but generally they 
have introduced faults greater than the one they were 
intended to prevent. 

119. Cole's Bepellent Escapement. —Mr. Cole was 
the inventor of another form of lever escapement, which 
he termed the Repellent, or Anti-detached. The pallets 
on this escapement differ from the ordinary ones only in 
having a recoil rest instead of a " draw ; " but the lever 
and roller action is different. The lever is brought up 
to a narrow point, instead of having a forked end, and 
gives the impulse through a small notch cut in a small ruby 
roller against which it rests, while the balance is perform- 
ing what would be its free arc of vibration in the ordinary 
lever, being gently pressed by the repellent action of the 
wheel tooth on the locking, or resting, face of the pallets. 
Like the previous escapement, this one requires no bank- 

1 I 


ing pins. It approximates to the action of the Duplex, to 
which it is much inferior, because of requiring a larger 
roller to obtain sufficient impulse, and for other obvious 
reasons. It has never been generally made. 

120. The Club-tooth Lever.— The only other lever 
escapement in general use, besides the table and double 
roller escapements, is what is called the "club-tooth'* 
lever. In this escapement the lever and roller action 
is the same as in the two former, but it differs in the 
pallet action, the impulse planes being partly on the 
teeth and partly on the pallet. It is almogt universally 
used in French and Swiss-made lever watches. This 
form of wheel permits of very little drop being allowed 
for between the escaping tooth and the next one coming 
into action on the opposite pallet face (the tooth being 
cut away at the back from beneath the impulse plane, 
to free the pallet comer when it is driven into inter- 
section with the wheel) ; whereas, with the taper-pointed 
teeth of the ratchet wheel, a small amount of space 
must be lost. It is also, from its shape, better calcu- 
lated to retain the oil. But these advantages are more 
than counterbalanced by the disadvantages attending 
the action of the two planes in contact, whose surfaces, 
at one part of the action, are nearly touching one another 
the whole length of the plane on the tooth ; so that, apart 
from the question of capillarity, any thickening of the 
oil — the necessity for which to the wheel teeth constitutes 
the great drawback to the lever escapement — ^will entirely 
upset the performance of the watch. 

With the sharp points of the brass ratchet-shaped 
teeth, the points slide down the planes with very little 
resistance, the action being so light that I have seen 
many wheels that showed no sign of wear after having 
been more than thirty years in action. 

I have heard escapement makers remark that the 
ratchet wheel is preferable to the club tooth, because 
there is less friction ; and this is, to a certain extent^ 
true, as the planes diverge more raj^ldl^ from each other, 

Fig. 31.— 8<nlgbt-Una Lever EKHpementt with Club Teetb. 


and the angular motion of the part of the tooth in contact 
increases as it passes over the pallet plane. But it is a 
mistake to suppose that the extent of the surfaces in 
contact makes any difference, for anyone with the most 
elementary knowledge of mechanics knows that the 
friction would be the same in both instances, other con- 
ditions being constant ; indeed, friction has very little to 
do with the retardation of such light bodies, as compared 
with thick oil, badly-formed parts, &c. 

121. Straight-line Escapement. — The Swiss attach 
great importance to what is called a straight-line escape- 
ment (Fig. 34), which is never made in England, as, with 
our fusee movements, it would be difficult to find room 
for it. It is supposed that there is less friction and shake 
on the pivots from this arrangement, from the direction 
of the pressures neutralising each other to some extent ; 
but there is really no advantage to be derived from plant- 
ing the pallet at any one angle more than another (the 
];)arts being detached from one another after the action 
has taken place), unless, perhaps, it be in the look of it, 
with its visible stones and fancifully-shaped lever, an 
advantage to which the Swiss are not insensible. 

122. Pallets with Circular and Equidistant Lock- 
ings. — Adifference of opinion exists amongst watchmakers 
as to the relative advantages of circular pallets and 
pallets having lockings, circular, or equidistant, from the 
centre of the pallet-staff. If the arms of the pallet are of 
equal length, the locking faces will be at unequal dis- 
taiices, and the resistance to the unlocking will be appa- 
rently different, owing to the greater leverage on the one 
than on the other, and therefore the pallets with equi- 
distant lockings have been recommended, and notably by 
Mr. Grossmann, who strongly advocates them in h^ 
admirable essay and elsewhere ; but they have never been 
adopted either here or in Switzerland. (Figs. 35, 36.) 

It is found practically that a properly-constructed 
escapement will perform equally well with either pallets 
— with equidistant lockings or circular pallets — so that 


really there ia no reason for changing the latter for the 
former, especially as the making of the circular pallets ia 
EM syatematised and well understood by pallet makers; 

Tie. 35.- ClnsDlaT Falleta. 

but for the benefit of those who are fond of analysing 
theories, I will endeavour to summarise the a:^;uments 
for both sides of the question. 

Taking first those in favour of equidistant lockings, 
it can easily be proved geometrically ttiat the friction on 
the impulse planes of pallets moving through the same 
lifting angle, where the impulse is delivered tangentially, 
is the same whatever may be the length of tJhe pallet 


arm or its distance from the centre of motion. There 
is, therefore, no advantage, tlieoretically, in having tlie 
impulse planes circular, or at equal distances from the 


FaUeta with Eqnidliitant LoaUngs. 

centre. Whereas it is apparently obvious that with the 
locking faces at unequal distances from the centre, as in 
circular pallets, there is a greater amount of resistance to 
the unlocking on the entrance pallet than on the exit. 
This would be tme if the locking faces were circular, as 
in the Graham "dead-beat" escapement ; but as there must 
bea certain amount of "draw "on the locking faces of lever 
pallets for safety, the greater excuiimt lA tcfSMOL ticwa ^ven 

Chap. XIV.] 



with the one pallet than with the other causes a greater 
amount of resistance to the unlocking on that face ; 
while the necessity of using oil on the impulse faces 
considerably modifies the first part of the proposition, in 
favour of equidistant impulses, as it is obvious that with 
a longer and more oblique impulse plane the oil will 
have a greater retarding effect than with a shorter. 

It will be seen by consulting the diagram (Fig. 37) 

Pig. 37. 

that the unequal resistance to the unlocking in circular 
pallets is more apparent than real. Let a and b 
represent the flukes of a pallet having equidistant lock- 
ings ; n V rriy z t X are arcs of a circle, struck from the 
centre. A, through which the locking points will pass; 
and whatever depth the tooth locks, it will be made to 
.i recoil to those arcs, as the comers t, v of the locking 
faces pass over them when they are drawn out from the 
wheel; p r and h 8 are the angles (12°) of "draw " on 
the pallet, taken off on the radii h t o, p v o, from the 
points t and v. 

It will be seen that, whereas on tSie \^i\Av».TA \»ft^ "^^ 


distance the tooth will be made to recoil is only the 
amount the choid z t deviates from the face « t, on the 
right-hand face the amount of the recoil is 12° plut the 

deviation of the chord n v from the radial line p v. It is 
evident, therefore, from the above proposition that tlie 
amount of reaistance to the unlocking is greater on the 
lockin;^ face of h than on that of a, and this may be 
equalised by either allowing a greater angle of "draw" 
oa a than on b, or by shortening the lever k. v. A good 
inaji^ escapements are maile'*iit^'«raLioa(!'iBsSBi.\atfi, ie., 


pallets having neither the lockings nor the impulse planes 
equidistant from the centre, but between the two, as 
shown in Fig. 38. 

123. Caliber. — On account of the difference of 
opinion of manufacturers as to the coiTect proportions 
and sizes of escapements no fixed calibers have been 
determined upon, some manufacturers preferring them 
larger and some smaller, &c., owing probably to the fact 
that they have been found to perform equally well when 
constructed with widely varying proportions. And as 
this difference of opinion extends also, as I have before 
observed, to the calibers of the movements, it is neces- 
sary that the escapement maker should understand how 
to adapt his escapement to these various conditions.* 

This want of system, though bad in itself, has tended in 
its results to develop the true theory of the lever escape- 
ment in England, where it is generally better understood 
than in other countries. Although pallet making is a 
distinct blanch of the business, the escapement maker 
should be able to make pallets for himself, and to measure 
that the angles are what they are represented to be. 

The size of the escape wheel determines the size of 
the pallets, which are the groundwork of the caliber of 
the escapement, and this should have some relation to 
the size and weight of the balance, which should be 
determined with reference to the size and depth of the 
barrel, and, consequently, the strength of the mainspring. 
With a balance of the correct size, a wheel of from three- 
sevenths to three- eighths its diameter will be about right. 

The numbers denoting the sizes of escapements have 
no reference to the sizes of |-plabe movements, but 
are taken from the sizes of movements to which they 
were applied when full plate watches only were made. 
When |-plate watches were introduced, the room for the 

* The woid " caliber ** is applied technically in the watch trade to 
denote certain conditions, proportions of the parts, &c., and not in 
its more limited and general sense : thus we speak of a movement of 
left-handed caliber^ a right-handed caliber, &c 


escapement was much more restricted, smaller and 
heavier balances were used, and the superiority of smaller 
escapements was demonstrated. The old numbers, 
however, are still retained, although they do not repre- 
sent anything, so that what is known as a four-size 
escapement will be the size for a ten or twelve-size 
movement (J-plate), and a six-size escapement for a 
fourteen or sixteen-size movement. 

The wheel and pallets being obtained, the frame is 
prepared for jewelling. 

124. Steady-pins. — The steady-pins in the cocks 
are usually put in by the movement maker; but as 
these pins, and the holes to which they are fitted, are 
rarely concentric with one another, and the pins are 
made of straight wii-e screwed into the cocks, while the 
holes are opened with a taper broach, they never fit 
well. Now, as a badly steady-pinned cock will be 
always liable to shift, to put the arbors out of upright, 
and so to destroy the very nice action of a good escape 
ment, the escapement maker has always to resteady-pin 
the cock himself. Tliere is no reason why the move- 
ment maker should not do this work as well as anybody, 
and no doubt he would, if manufacturers would only 
agree as to the exact calibers and positions of the parts. 

The escape cock is often left by the movement maker 
without pins, for the convenience of being shifted for a 
larger or smaller escapement. 

If the cocks are not steady-pinned, the soles should 
be rubbed flat on glass with oilstone dust and oil, and a 
small square notch filed or cut in their outside edge 
slanting towards the screw hole, to admit the point of 
strong tweezers for lifting the cock off the plate. 

The holes for the pins should be marked as far apart 
as the space will permit, and drilled in the drilling tool, 
with the cock screwed on to the plate in the proper 
position. The holes and pins should be drilled and 
fitted one at a time : the holes broached from the cock, 
and the pins filed in a chuck m a \aitlifi or throw, and not 


in a hand-vice, to the same taper. The pins must be 
short and well-rounded at the points, and driven into the 
cocks. The brass of which they are made should be of 
the same quality and hardness as the cock, as if it be 
much harder the pin will expand in the gilding, and 
show. If the pins are already in, they should be re- 
moved one at a time, and the holes broached, new pins 
being fitted as before. Cocks steady-pinned in this way 
will be perfectly tight and immovable when screwed into 
their places, although easily removed ; whereas with 
straight pins they are never steady, and the difficulty of 
removing them is a permanent danger to the pivots. 
There is no cause of bad performance in a watch more 
frequent than injured balance staff pivots, through the 
great difficulty jobbers find in taking the cock off and 
putting it in its place again. 

125. Sinking the Escape Wheel. — The space be- 
tween the pillar plate and the lower face of the escape 
cock may be determined by the thickness of the pallets 
and the lever, as there will be sufficient freedom for 
the fourth and escape wheels, if there is room enough for 
the pallets and lever. 

In flat watches the escapement is sometimes sunk in 
the pillar plate. The only difference in making the 
escapement in this way is that the escape wheel is run in 
below the fourth wheel, and the lever is placed above 
instead of below the pallets. 

There is a great advantage in this disposition where 
there is little room, as the distance between the escape 
cock and the pallets prevents the latter from drawing 
away the oil from the hole, which frequently happens 
when the pallets are close to the cock and too much oil 
is applied, especially if the holes are brass ones. 

If the escape wheel is sunk in the plate, the sink 
should be bevelled off at its edge, to permit of the wheel 
being removed when the watch is together ; or if the 
wheel is low in the plate, a slot cut between the two 
sinks will be better. 


If the escape cock is to have thorough holes, it should 
not be less than 02 of an inch in thickness ; and if it is 
to be jewelled with end-stones, not less than '027 of an 
inch. It should be screwed into its place, and turned 
down in the mandrel to the required thickness, so that it 
may be parallel to the pillar plate. 

The width of the balance lim will determine the 
space between the escape and balance cocks, but nothing 
is gained by keeping the balance cock above the plane of 
the top plate, and therefore it is best to turn out the 
freedom for the balance after the cock is jewelled, which 
is done by cementing the cock on to a piece of brass, and 
turning out the freedom for the rim and screw beads in 
the mandrel. 

126. Balance. — Assuming that th^ dial is made 
and that the seconds hole in it has to be worked to, 
and also that the escape cock is planted in a position 
suitable to the size of the escapement to be used ; 
if the dial hole does not exactly coincide with that 
in the frame, the latter must be stopped and another 
one marked from the dial hola This may either 
be drilled or marked sufficiently to work from. By 
placing the fourth wheel and the escape pinion in 
the depthing tool, the depth can be ascertained and 
marked on the bar; the distance may be taken from 
the balance hole to the centre of the escape cock, and a 
segment of a circle, cutting the sc )re already made for 
the depth, will give the exact position for the escape 
wheel j and when those holes are drilled, the frame is 
ready to have the balance staff and escape wheel holes 
jewelled. When jewelled, the balance steff and escape 
wheel pinion can be pivoted and run into their boles. 
The balance staff should be all steel, and as hard as it will 
turn. The proper height for the balance is ascertained 
by shortening the staff before the pivots are made until 
the balance is close to the escape cock, as the pivot 
cannot be shortened much when made. 

The balance should ^t vexy t\^V\y otl \3afe ^oat made 


for it on the staff, very little of which sliould be left for 
riveting; and the shoulder against which it is riveted 
must be square, not undercut. If this is not attended to 
the balance will not run true when riveted, and ad- 
ditional riveting will only make it worse. The seat of 
the balance spring collet should be polished and only 
tapered a little. The height for the total length of the 
staff can be got by taking out the jewel holes, screwing 
the end-stones in their places, and filing a piece of brass 
to the height between them, allowing of no shake ; or a 
still better plan is to take off the end stones and measure 
the distance between the outside of the holes with a 
douzieme tool. The escape wheel is fitted to a brass 
collet, or a gold one is better, driven on to the pinion ; 
the height of the wheel being easily ascertained, as it is 
close to the escape cock. 

The seat of the collet should fit the wheel very well, 
and here, also, very little must be left for riveting, or 
rubbing, the wheel on to it. Before fitting, the wheel 
should be put on a true arbor, and, if not true, the hole 
should be drawn until it is so, and broached round and 
fitted to the collet. 

The roller may now be roughed-out and let on to the 
staff. The position for the pallet holes being marked ofl^ 
the distance between them and the balance holes is tech- 
nically called the chord of leverage, and the escaping arc 
is entirely dependent on the way in which this chord is 
divided between the roller and the lever, i.e., the acting 
diameter of the roller, the acting length of the lever, the 
lifting arc of the lever (or pallets), the lifting arc of the 

If the pallets are 10^ pallets — 1.6., giving 10° total lift 
on the lever each way — and the balance is required to 
have a lift of 30? (a proportion very generally adopted), 
the chord of leverage must be divided into four equal 
parts, and the impulse pin in the roller placed at one of 
these parts from its centre, leaving the remaining three 
for the effective length of the lever. If the imigulsa ^vc^ 


is at a greater relative distance from the centre of the 
roller (in a table roller escapement) and the escaping arc 
consequently diminished (see page 140 and Fig. 16) the 
guard pin would not intersect the roller sufficiently to 
have a safe action ; and, therefore, when a short lift at 
the balance is required a double roller escapement is 
necessary — a larger roller to receive the impulse and a 
smaller to insure a safe action. 

127. The Impulse Pin. — The impulse pin should 
have about one-third of its circumference flattened 
off one side (or it may be made a segment of the 
arc described by the pin, which will be nearly flat 
in so small a space), and, the distance of the pin 
from the centre being determined, it should be marked 
on the roller, a hole drilled and broached to not quite 
the diameter of the pin, and a punch of the shape 
of the pin put into the hole with the flat side towards 
the edge of the roller, the edge of which should be 
beaten in with the psean of a small hammer until the 
hole takes the form of the pin. If the punch be a little 
taper it can be driven into the hole until the pin fits it. 
The pin should fit well and not be dependent upon the 
shellac with which it is fixed for its stability, and it 
must be perfectly upright. 

The roller must be turned to the proper size, and the 
small crescent in front of the impulse pin filed out to 
permit the guard pin to pass when the lever is giving im- 
pulse. There is no method for determining the outer 
diameter of the roller, but, with an escaping arc of 30® at 
the balance, the intersection of the guard pin with it will 
be very small indeed, and, as this pin must be quite close 
to the roller when the impulse pin leaves the lever notch, 
the roller must be turned down as small as possible. 

The position of the pin may be determined, with any 
lift, by drawing a line x b (Fig. 39) at half the angle of 
lift on the lever to a horizontal line A B, which represents 
the line of centres. Placing one point of the depthing 
tool in the hole or marked point B, and the other in the 




centre of the roller at c, at the distance of the respective 
centres apart, the point of intersection of the roller with 
that line, as at x, x', x'\ will give the exact point for the 
guard pin, whose thickness must then be allowed for. 

With a one-to-three escapement, as it is called — i.e., 
one having one jmrt of the distance on the line of centres 
taken up by the radius of the roller, and three parts by 
the acting length of the lever — the intersection is found 
to be half the thickness of the pin ; and a very good 
method of determining the position of the pin, adopted by 

Fig. 89. 

escapement makers, is to draw on a piece of smooth 
brass a series of arcs from the centre of the lever, and 
placing one point of the depthing tool in that centre and 
the other in the centre of the roller, the arc which its cir- 
cumference touches will be the distance from the centre 
of the hole at which to drill the hole for the pin, so that 
half its thickness will intersect the roller. 

Before drilling the hole for the pin or the pallet staff 
in the lever it is better to file up the lever fork and cut 
the notch, and for this reason : the notch must fit the 
impulse pin without shake and be cut quite square to 
the bottom, as there is no room to spare, and the guard 
pin must be exactly in the middle, and it is much easier 
to mark and drill l^e hole for the guard pin than it is to 
file the notch to the hole when that is drilled first. 

128. The Guard Pin.— The hole for the guard 


pin can be marked and drilled as near to the bottom 
of the notch as possible, and the hole for the staff 
drilled from it. The lever may be roughly shaped, 
and, if the hole in the pallets is quite square to 
the sides, or upright, the pallets and lever should be 
held together in the spring tongs and broached for 
the pallet staff; it is well to firet put the pallets on 
an arbor in the turns, and try, by touching the sides 
with a graver, if they stand square to it ; if not, the hole 
must be broached slantingly until it is square to the 
sides. Notwithstanding the numerous templets and 
gauges used for the measurement of the various parts 
and proportions, it is quite impossible to insure that 
absolute exactness requisite for the j^erfect working of 
an escapement, so that it is always necessary to examine 
it with the parts together when complete. The roller 
and lever depth is easily seen in the tool, and can be 
examined when the lever is let on to the staff, the dis- 
tance on the plate being taken with the points of the 
depth ing tool on the dial side of the plate and marked 
from the balance jewel hole with the end-stone removed. 

The depth of the wheel and pallets can be marked 
in the same way ; and this cross depth requires to be 
very accurately determined and the hole carefully drilled, 
the two actions of the roller and lever and the wheel and 
pallets being dependent upon its position- The lighter 
the marks made with the depthing tool, and the smaller 
the dot or centre to drill from the better, as the smallest 
deviation from the correct spot will necessitate re- 

129. The Pallet Staff.— When the pallet staff* has 
been run in, the depths can be tried in the frame, and, 
as the action of the wheel and pallets is more easily 
seen in the frame than in the tool, it should be tried 
first. If the lockings are safe and the wheel will 
escape forwards, the freedom can be tried by running 
the wheel backwards, pressing it towards the pallets; 
if it runs freely the dept\i ^VW \i^ ^ovrect, but if in 


going forward the tooth falls on the plane of the 
pallet, or even on the comer, it will not be safe, and the 
pallet staff must be replanted, as would be the case where 
the depth is much too great. Topping the wheel must 
not be resorted to. 

130. Lever and Pallet& — ^The angle of the pallets 
with the lever may be obtained approximately by 
moving the wheel forward, and observing if the lever 
falls more on one side of the balance holes than on 
the other. The balance should then be put into 
its place, and if the guard pin passes the crescent, 
it shoiild just touch the comer as it passes out, the 
polishing of the edge of the roller giving it suffi- 
cient freedom afterwards ; and if it stands close to the 
edge of the roller on either side, when a tooth escapes 
and the lever has passed the crescent, the escapement 
— i.e., the lever and pallets — will be what is termed in 
angle, and the banking pins can be marked while the 
guard is pressing the roller. The necessary freedom for 
the guard pin from the roller may be got by either 
making the roller a little smaller, or the lever a little 
narrower, where it rests on the banking pins. The 
position of these pins should always be near the fork of 
the lever; they should be small, and, if left no longer 
than the height of the lever from the plate^ there will be 
little danger of their being bent. 

If, as is sometimes the case, the lever is just a shade 
too deep or too shallow on the roller, the hole for the 
staff may be " drawn " a little. This is done by gripping 
the lever in the spring tongs or pliers (with brass faces), 
coveiing a small portion of the hole on the side from 
which it is to be drawn. By broaching the hole while 
held in this manner you make it a little oblong, and it 
must be punched up a little on the opposite side, as it 
must fit tightly on the staff 

After both wheel and roller depths have been made 
correct, the pallets and lever must be clamped or held 
together, and the holes drilled on each sid^ oi '<}cl^ ^^»& 


for the pins that are to hold them together. The lever 
must be filed up, hardened, and tempered. The lighter 
it is made the better, and it should be made of such 
a shape that it will not cover the pallet arms and draw 
away the oil from the wheel teeth. The pallets and 
lever must be carefully poised ; if the pallets are to be 
made thinner befoi*e polishing, they should be put on an 
arbor in the turns, and have races cut in the sides with 
a graver, and the sides filed down to these races. A 
good deep sink shoidd also be cut in the pallets in front of 
the staff hole, to prevent them from drawing the oil away 
from it The back and inside of the pallets are polished 
in a swing tool, and the sides on a cork ; care must be 
taken that too much pressure is not used and the stones 
loosened, as a loose stone is a grave fault and not easily 
discovered. It is best to use a steel polisher with oil- 
stone dust at first, finishing with a zinc polisher and 

131. Escape Wheel. — The escape wheel should be 
put on an arbor, and the rim turned true on each 
side to the required thickness, and the arms and centre 
reduced by a piece of blue stone, as, if either the 
escape wheel or the pallets are out of poise, it will 
affect the going of the watch. When the wheel 
is smooth and flat on the sides, it should be polished ; 
no escape wheel should be gilded. The way to polish 
a wheel is as follows : — Take a piece of brass, not 
much larger than the wheel, rub it on glass with 
oilstone dust until it is quite flat ; make a small hole 
in the centre, and a few circles on it about the size 
of the wheel, and with a cement made of bees-wax and 
a small grain of resin, cement the wheel on the block as 
near the centre as possible ; the circles on it being a 
guide. The wheel may be polished underhand on a tin 
block filed flat and scraped very clean, by placing the 
point of a peg in the hole in the piece of brass. Care 
must be taken that the pressure is equal, as it is very 
easy to get one side a little more rubbed than the other. 


One side having been polished, a very little heat will melt 
the cement, when the wheel can be reversed, and the op- 
posite side polished in the same way. The cement can be 
removed from the wheel by heating it in a little clean 
oil until it is melted, and what still adheres will be 
easily breaded off. 

The polishing stuff used for the wheel should be red 
stuff (not too fine) as diamantine will not polish brass. 

When the wheel has been polished it is ready to be 
rubbed on to the collet on its pinion, after which the tops 
of the teeth should be touched true by holding a slip of oil- 
stone against them, and rotating the wheel in the turns. 
The roller has generally a hollow cut out of it for 
ornament in the centre, and its face polished on a block. 
The edge, which is left square, is polished by bringing it 
into contact with a revolving disc in the lathe, or in the 
turns. The Swiss make the edge of their rollers round, 
and in the older Swiss watches they are often found with 
it tapered off to an angle, in order, it is thought, to avoid 
friction, in case the guard pin should come in contact 
with it. But English escapement makers disbelieve in an 
escapement that would permit of this frequently occur- 
ring, and the guard pin is used merely as a precaution 
against a remote contingency. 



132. Invention. — Perhaps the most remarkable epoch 
in horological chronology was marked by the invention 
of the balance spring ; its application to pocket watches 
placing within the reach of everybody the means of 
regulating and registering the daily actions, «fec., with a 
degree of exactitude quite unattainable with the time- 
keepers hitherto made. 

This invention is generally accredited by foreign 


writers to Huygens, who applied for a patent for it in 
Paris in 1674, in which he was opposed by L'Abb^ 
Hautefeuille, who proved that he himself had applied for 
the patent some years before. Although Dr. Hooke, 
whose invention it undoubtedly was, had applied for 
a patent for it here sixteen years before that time, both 
of these inventions may possibly have been original, 
so far as the inventors themselves were concerned. 
Hooke himself disbelieved this, and he accused the 
Secretary of the Royal Society (Mr. Oldenburgh) of 
making known to foreigners the inventions deposited in 
the records of the society. He did not complete his 
patent, owing to a disagreement with certain parties 
with whom he was to have taken it out conjorutly (and 
who afterwards negotiated with Huygens on the same 
subject), and allowed it to fall through. 

133. Hookers Law. — It is to Hooke we owe the 
exposition of the qualities of the spring (both as ap- 
plied to watches and otherwise), as it is to him we 
owe its application; the law of elasticity enunciated 
by him (see page 18) being an axiom in mechanics 
with which everybody is familiar. But this law is only 
partially true as applied to balance springs of watches 
and chronometers, otherwise every spring would be 
isochronous. Pierre Le Roy was the first to publish this 
discovery, although it is probable, from Hooke's genius 
and the many experiments he made on springs, that he 
had found this out also. Le Roy says that " there is in 
every spring of a sufficient extent, a certain length where 
all the vibrations, long or short, great or small, are iso- 
chronous ; ** and that " this length being secured, if you 
shorten the spring, the great vibrations will be quicker 
than the small ones ; if, on the contrary, it is lengthened, 
the small arcs will be performed in less time than the 
great ones." 

134. The First Spiral Springs.— The first spiral 
springs that were applied to watches were hand- 
made — that iS) turned up by hand by means of a 


small steel peg or burnishing tool, which, by biimishing 
one side of the wire, caused it to coU up as a 
slip of paper will if scraped on one side with a knife. 
Some of these springs were very beautifully made, and 
were far superior to the soft tool-made springs after- 
wards in use, as the burnishing which turned them up 
also hardened them to a certain extent, and set them in 
spiral form. As soon, however, as different forms and 
i lengths of the spring were introduced, it became 
necessary to coil them up on blocks, &c. ; and shortly 
afterwards, as the soft springs were found to lose their 
elasticity, they were hardened and tempered, and con- 
tinue to be made so for all high-class watches. Forty 
years ago every finisher made and applied a spring to the 
watch he finished, except to a few of the best watches, 
which were sprung with the above-mentioned hand-made 
springs, or the ordinary tool-made springs hardened 
between two pieces of brass screwed together — ^a very 
clumsy and almost impossible way of hardening a spring 
and keeping it in shape. The springs of forty years 
ago were turned up in a tool, and blued to give them 
a set; now, very few finishers know anything about 
either making or applying them. 

135. Different Forms of Spring. — M. Lutz, of 
Geneva, discovered a method of hardening springs 
other than the old process of fire and water, and of 
imparting a very beautiful appearance to them. He 
exhibited a case of these springs in the Exhibition of 
1 85 1, which afterwards came into my possession, and very 
charming things they were, in all forms and colours. I 
believe the Commissioners gave him some assistance to 
enable him to perfect his invention, but I do not think 
these springs have been improved upon since then, al- 
though the process is still a secret. Great quantities of 
these springs and imitations of them have come into this 
country, superseding the oM tool spring of the finisher, 
as they deserved to do, but they are altogether unworthy 
of a good English watch, although too often applied 


to watchen receiving that appellation, as they have none 
of the properties of a properly hardened and tempered 
spring. Many different forms of spring (Fig. 40) have 
been in use from time to time, the most successful of 
which are the Breguet spring, invented by the French 
watchmaker of that name, and the cylindrical helical, 
which was the invention of John Arnold the elder, who 
patented it in 1775. The latter of these springs was in 
general use for all the higher class of pocket watches 
until comparatively recently, but since the qualities of 



Fig. 40.— 1. Ordinary Volute. 2. Cylindrical Helical. 3. Bregaet SpEing; 

the Breguet spring, which is better adapted for the 
thinner watches now made, have become understood, it 
has quite superseded its rival in all but pocket and 
marine chronometers. 

136. Isochronism of Balance Spring. — A balance 

spring, of whatever form, to be isochronous must satisfy 
the following conditions : — Its centre of gravity must 
always be on the axis of the balance, and it must expand 
and contract in the vibrations concentrically with that 
axis. When these conditions are secured in a properly 
made spring it will possess the quality of isochronism— 
that is, its force will increase in proportion to the tension, 
and it will not exert any lateral pressure on the pivots. 
M. Phillips, in his memoir, demonsti'ates these condi- 
tions, and proves theoretically that the terminal curves 
deduced with the view of satisfying the one condition 
verify at the same time the other. 

137. Isochronism and Length of Spring.— There 
is a theory, attributed to the late Mr. Charles Frodsham, 


that every length of spring has its isochronous point; 
and Mr. Immisch, in his prize essay, says : — " Every 
one with some experience of timing knows that mere 
length has nothing to do with isochronism." In a paper 
on the subject which I contributed to the Horological 
Journal of June, 1875, I denied this, and asserted that 
length had everything to do with it— that a spring too 
short, whatever its form, would make the short arcs of 
the balance vibration be performed in a less time than 
the long arcs, and a spring too long would have just the 
contrary effect, and for this reason. If you take a balance 
in its place in the watch with a spring of ten turns, a 
length often prescribed, and move the balance round half 
a turn, you will find that, while the inner end of the 
spring, being pinned to the collet, moves round in the 
same circle as the balance, the outer end being fixed, the 
disturbance in the spring taking place in the inner coil, 
has gradually affected the fixed and rigid end, and, in a 
spring of this length— especially if it is a hardened spring, 
and therefore more elastic— the maximum resistance has' 
been reached, or nearly so. Therefore, if you move the 
balance another quarter of a turn in the same direction, 
you have added little or nothing to the resistance, and the 
spring has not force enough to throw the balance back 
through the longer arcs in the same time that it would 
have gone through the shorter. A spring with ten turns 
is, therefore, too short, and in consequence a watch with a 
spring of that length will gain in the short and lose 
in the long arcs. Now, take a spring of twenty turns and 
go through the same operation, aiid you will find that the 
balance must be moved through more than the half-turn 
to disturb the fixed end of the spring ; but as the spring 
is more than double the length, it is also thick in propor- 
tion, and when the disturbance reaches the rigid end, it 
exerts a force sufficient to carry the balance through the 
longer arc in a shorter period than before that force was 
called into action. A flat spring with ten turns is, then, 
too short, and one with twenty too long ; for a spring 


with ten turns will make the watch go fast in the short 
arcs', and a spring with twenty turns will make it go slow 
in the short arcs. 

138. Compensation. — The compensation is not much 
affected by any difference in the lengths of the springs, 
there being a compensating power in their expansion and 
contraction by which there is a gain in favour of the 
spring, so far as concerns its dimensions, and it is only 
the difference in the loss of elasticity between the two 
which has to be compensated for, as was pointed out 
recently by Mr. Wright, the lecturer at the Horological 
Institute. This is in accordance with the law that the 
resisting power of a spring varies as the cube of its 
thickness, and as the spring expands equally all ways, it 
would become actually stronger with any increase of 
temperature, were the elasticity unaffected by the change 
which takes place in the molecules, and this difference of 
elasticity in springs of different lengths is very small. 

My friend Mr. Mercer, who, from his large experience 
in chronometer timing, may be considered an authority 
on the subject, tells me that in many instances he has 
not had to touch or alter the weights on changing the 
springs, and that, at all events, it is not certain which 
way the alteration of the compensation would be re- 

139. Flat Springs.— I find the best length for a flat 
spring to be fourteen turns, and if it is half the size of 
the balance, and pinned in properly, the time will be 
pretty near in positions. But although the flat spring 
is the most common, it is also the worst form of spring. 
It cannot expand on the side next to the stud, but has a 
dragging lateral action, while it throws out on the oppo- 
site side to a third more than its proper size, causing 
considerable pressure of the balance pivots on each side 
of the holes alternately. 

It will assist the action of this spring if it is always 
a little small, as this gives more freedom to that portion 
of the coils next the stud. 


140. Correcting the Spring. — I have seen directions 
by experienced men how to time in positions by the curb 
pins. This should never be attempted. 

The curb pins (always an evil) should be wide enough 
apart to let the spring just move between them, and no 
more, and should never be far from the stud. As 
" manipulating " the curb pins, as it is termed, is done 
only with the object of lengthening or shortening the 
acting length of the spring, this should be accomplished 
in the proper way at once, by adding to or taking away 
weight from the balance. Should the watch with the 
spring referred to gain a few seconds in the short arcs, 
taking up the spring half the width of the stud, and re- 
placing two of the balance screws with heavier ones, 
will remedy this defect. However, should it do the reverse 
— that is, lose in the short arcs — by taking a very little 
off the weight of the balance, and letting the spring out 
so much, the error will be corrected — if the error is in the 
spring — but no directions will enable any one to make a 
watch go well with bad pivots, bad holes, and large and 
bad escapements. The Breguet spring, although differ- 
ing very little in form from the simple volute, is essen- 
tially different in action and principle; the overcoil, 
being fixed above the spring, and nearer the centre, gives 
it perfect freedom to expand in a circle all round. This 
spring must be longer than the flat spring, as the force 
of the outer or fixed end is sooner reached, and the 
curve inwards gives it more power of resistance, and also 
an easy and perfect means of obtaining isochronism. I 
find about twenty, according to the size of the watch, the 
best length for this spring, and curb pins should never be 
used with it, if perfect timekeeping be aimed at. 

I would also warn any one against altering the 
shape of the balance staff pivots ; there is only one shape 
for these pivots, and spoiling the pivots is not iso- 
chronising the spring. This remedy is almost as bad 
as putting the balance out of poise to obtain time in 
positions. My theory is length, and length alone, and 


this seems to me to be in perfect accord with the result 
of a good many experiments on this subject, most of 
them pursued unconsciously ; as, for instance, the taper- 
ing of the wire before making the spring, and, again, in 
the contrary direction, when what was termed the iso- 
chronous stud — that is, a long spring stud — was used. 
I have actually seen these two remedies applied to 
one watch, when the result desired might have been 
obtained by a spring of the proper length The difficulty 
of getting the short arcs sufficiently fast is much in- 
creased in small watches, owing to the comparative 
thickness of the staff pivots, while in marine chrono- 
meters the trouble is usually to get them slow enough, 
the pivots being relatively smaller ; but this difference 
may not be entirely due to the pivots, the conditions 
under which the long and short arcs are tried in chrono- 
meters and watches being different. Mr. Crisp tells me 
that the late Mr. Charles Frodsham had his marine 
chronometers timed in positions, and that in some of 
those he timed for him he found the tendency to 
lose in the short arcs the same as in watches, proving 
that the side friction on the pivots has a good deal to do 
with it. Various materials have from time to time been 
tried with a view to the prevention and lessening of 
the compensation error. Tlie first Mr. Dent had some 
ordinary steel springs electro-gilded, but, as might have 
been expected from the different natures of the metals, 
such a combination did not answer. The gold did not 
adhere weU to the steel, but peeled off in places, possibly 
owing to the different expansion of the metals. Glass 
springs have also been tried by more than one maker, 
but with little success. Mr. Dent made a number of 
experiments on )hem, but with no other result than to 
demonstrate that they were quite impracticabla There 
was one of these springs in a chronometer (going) at the 
Exhibition of Scientific Apparatus at South Kensington 
(1876), of which a member of the firm of E. Dent and 
Co., speaking at the Conference, said: — **This glass 


spring is the invention and handiwork of the late 
Frederick Dent, and the chronometer before us represents 
the only perfect specimen now in existence. The elas- 
ticity and strength of this spring are so extraordinaiy 
that until you had examined it you could scarcely credit 
its being glass But meanwhile, for illus- 
tration, we will tell you two experiments, the second 
one being quite unintentionaL In the first, a chrono- 
meter wa^ taken, having an ordinary steel spring, the 
balance of which was oscillating 180** from rest. A glass 
spring was substituted, everything else remaining exactly 
the same. The oscillation of the balance rose to 200'." 
The unintentional experiment was that the chronometer 
to which this spring was attached was accidentally 
knocked off a table, and, although the staff pivots were 
broken, the spring sustained no injury. 

The same gentleman said this chronometer had always 
a tendency to accelerate or advance upon its rate, and 
referring to some experiments on chronometers with 
springs of different materials, with glass balances, he said, 
it was found that, while one with an ordinary steel spring 
lost six minutes twenty-five seconds a day in a rise of 
68^ Fahr., the one with a glass spring lost forty seconds 

141. Glass Springs — If glass could be made to 
answer as a material for balance springs, it would be a 
matter of the greatest importance to the trade. We should 
get rid of all our troubles with the compensation balance ; 
in fact, we should not require compensation balances at 
all for ordinary watches, glass undergoing only about a 
tenth part of the change for variations of temperature 
that steel does. The error of six minutes in twenty-four 
hours in a range of 60° Fahr., for which we have now 
to compensate, would be reduced to little more than 
half a minute, and all our disputes about secondary 
errors, and the best form of auxiliary compensation, set 
at rest. The discovery of a metal suitable for springs, 
and possessing the same amount of indifi'erence to 


thermal changes as glass, would create a revolution in 
the trade, and prove a fortune to the finder. Glass 
springs, owing to theii- want of elasticity, have to be 
made quite double the length of steel ones, and, from 
the great difficulty of making and applying them, have 
never been made, except for the sake of experiments 
like the foregoing ; and they could not for the former 
reason be applied to pocket watches, where the space is 
so limited. 

142. Palladium Springs. — Great advantages are 
claimed for the palladium springs introduced by M. 
Paillard, as they cannot become rusty, are non-mag- 
netisable, and maintain their elasticity better in varying 
temperatures than gold or steel ones. The compensation 
error is certainly less with a palladium spring than with 
a steel one, and immunity from rust is of great value, as 
the smallest speck of rust on the spring will destroy the 
rate of the best chronometer ; to obviate this, gold springs 
have sometimes been used. The late Mr. Jiirgensen, of 
Copenhagen and Locle, advocated gold springs, and, some 
years ago, sent two ships' chronometers to Greenwich for 
trial so sprung, but the result was not satisfactory, as 
they stood low on the list. The specific gravity of gold 
is greater than that of steel, and its loss of elasticity 
greater in a given range of temperature, and, as gold 
springs must be formed by an annealing process, they are 
inferior even to the hard-drawn steel springs that were 
made before fire-tempering was introduced, in every re- 
spect, except their unmagnetic and non-rusting qualities. 

The loss of elasticity in a given range of temperature 
being less in palladium than in steel springs, and the so- 
called middle temperature error being consequently re- 
duced, has induced several chronometer makers to apply 
them to chronometers intended for the Greenwich compe- 
tition, and chronometers with ordinary balances (without 
auxiliaries) have held a veiy good place in these trials. 
JBub, while chronometer makers should pay every attention 
to attempts to obtain a \>et\feT xsv^Xfe-mX Iq^t the balance 


spring, it would be rash indeed to discard the well-proved 
hardened steel spring, with its known good quaHties, for 
any material, however promising, without very long and 
exhaustive tests of its qualities. The old soft tool and 
hand-made steel springs were found to lose their elastic 
force after the watch or chronometer to which they were 
applied had been a few years in use, and the instrument 
fell oflf from its rate in consequence, and, since there is no 
change in the molecules of a palladium spring to perma- 
nently set it, such as is undergone by a steel spring in the 
hardening process, it is possible that this spring may, 
after a few years, also lose its elasticity. Like all soft 
springs, it also requires very careful handling to avoid 
bending it so as to give it a permanent set in any but 
the right shape. 

As to its unmagnetic property, that would be of very 
little use where the ordinary compensation balance is 
used, as any magnetism which would affect the spring 
would assuredly reach the balance and equally interfere 
with the performance of the instrument, and any chro- 
nometer or watch made with the object of preventing 
this would require a balance, as well as the other quick- 
acting parts of the escapement, of gold or some other un- 
magnetic metaL 

143. Demagnetising Steel Work. — For the follow- 
ing description of an approved method of demagnetising 
portions of the steel work of watches which may have 
become magnetised, I am in debted to my friend, Mr. 
Latimer Clark. 

Watches sometimes become magnetised by being 
brought too near to a magnet or dynamo-machine ; in fact, 
persons often injure the performance of a watch by play- 
ing thoughtlessly with a common magnet. If the watch 
be taken to pieces, the separate parts may be demag- 
netised without much difficulty. If we taie any short 
piece of steel and approach one end very near to one 
of the poles of a magnet, it will become strongly mag- 
netised ; if we make the other end approach^ but uot ^j^xl^?. 


80 closely, the magnetism will be reversed in direction, 
and will at the same time be somewhat weaker than 
befora By repeating this process, always increasing our 
distance, we can rapidly remove all traces of magnetism. 
The simplest way to do this is to lay a small bar magnet 
on a divided rule, and holding the piece of steel in a 
brass or wooden pair of pliers, we approach one end 
within one-sixteenth of an inch, then the opposite end 
within one-eighth, then the other within three-sixteenths, 
and so on till we get to a distance at which the mag- 
netism disappears. 

We can readily test this by magnetising a portion of 
a very small sewing needle, and suspending it by a 
delicate fibre of silk within a glass tube or bottle. Our 
piece of steel if magnetised will attract this needle with 
one end and repel with the other ; if not magnetised both 
ends will attract similarly. 

The balance of a watch, being in the form of an S 
curve nearly closed, cannot be so easily treated, and the 
same is true of the spiral springs. To demagnetise these 
the best plan is to mount a horse-shoe magnet on a lathe 
so that its poles rotate in a vertical plane ; the poles 
should preferably be at least an inch apart, and the 
magnet should be well magnetised. The balance is sup- 
ported by means of a piece of cork on a small revolving 
turn-table which rotates in a horizontal plane ; this turn- 
table is canied on a wooden bar held in a slide-rest so 
that it can be slowly advanced up to the magnet and 
withdrawn. The turn-table is rotated by a kind of 
overhead motion fixed on the lathe table — this may be 
a round wooden rod with a pulley at one end driven 
from the lathe-head, while an india-rubber band traverses 
along it and drives the turn-table. Both the magnet and 
the table being set in rotation, the balance is brought up 
to the poles of the magnet by the slide-rest, and as nearly 
in contact with it as safety permits — it is then very 
slowly withdrawn. By this process every part of the 
balance is magnetised and deci'a.'gcist\aed in every con- 


ceivable direction, but the magnetism becomes weaker 
and weaker every moment, and ends by being so feeble 
that it is not injurious. The parts of a watch that 
require most attention are of course the balance and the 
pallets, but it is better to treat the balance spring also ; 
this last may be held on a cork by shellac solution and 
afterwards cleaned in alcohol. 

If we have command of the current from a dynamo- 
machine or from a very large battery so as to make a 
powerful electro-magnet, we may treat the watch in this 
fashion as a whole, without the necessity of taking the 
parts asunder. 



144. Testing Spring Wire. — Before setting about 
spring making the wire should be tested, and if it 
will not harden at a cherry red, be rejected at once, as it 
will require more heat to harden it when in the box or 
on the block. There is great difficulty in getting good 
wire for spring making, especially for small springs. The 
constant annealing required to make it soft enough to 
be drawn so fine decarbonises it, and makes it so mild 
that it will either not harden at all, or only at a heat 
that is utterly destructive to its qualities as a spring. 
As I recommended in a lecture delivered by me at the 
Horological Institute, in 1 877, the wire, when too mild, 
should be placed in a vessel filled with charcoal, and 
heated in a furnace until it is sufficiently recarbonised, 
in the same manner as steel was originally made, any 
difference in the relative hardness of the outside to the 
inside being rather advantageous than otherwise. If it 
is attempted to make springs of mild wire it will be 
found that the springs will stick to one another ul tlvA 



[Cliap. XVL 

winding-box when they are hardened, and the wire will 
scale, owing to the excessive heat to which it has been 
subjected, and will be very difficult to polish afterwards. 

145. To Make the Winder. — Take a piece of brass or 
German silver, about the eighth of an inch thick ; drill 
a hole in the centre, and turn it out in the mandrel to 
about f of an inch in diameter, in the form of a small 
watch barrel. This is called a spring-box. 

Fit a lid to it, also with a hole in the centre ; turn 
down a piece of brass wire with a pivot and shoulder to 
fit the small hole in the box, to pixjject a yery little way 
through it (it should project through the bottom of 1i» 

Pig. 41.— 1. Box -cover. 2. Spring-box. 3. Screw far fixing Wire. 4. Plan 

of Winder-pivot. 5. Winder, 

box the exact width of the wire to be used for the 

Tap a small hole in the centre of this winder-pivot, 
and cut three thin slits in a triangle to receive the ends 
of the wire of which the springs are to be made (Fig. 41, 
No. 4). Drill three holes at equal distances through the 
rim of the spring-box, close to the bottom, through 
which pass the wire, and fasten the ends with the screw 
(No. 3) into the slits in the winder. 

The head of the screw should project through the 
hole in the cover. Now the cover is to be put on, and 
gently pressed down during the process of winding the 
spring up, until the box is quite full, when the ends of 
the wire must be cut off, the screw removed from the 
centre, and the winder taken away. Bind the cover 
down with wire, stop the hole in the centre (animal 
charcoal and soap will do for stopping), and harden the 
whole in the usual way. Water is better than oil for 

Chap. XVI. 1 



thia purpose, and mercury Letter than either, as the lowei 
the temperature at which they are hardened, the bettei 
will the springs be. If the box is now put into a small 
vessel filled with oil (an old metal spoon will do), and held 
over a spirit lamp until the oil just catches the flame, and 
then left to cool, the springs will shake out of the box 
quite easily, and you have now three sjirings which, even 
in this state, are superior to the so-called hardened and 
tempered springs of Swiss make. If the air is effectually 
excluded, the process of hardening will not discolour the 
steel much, nor will the tem- 
pering in oil, and very little 
time ia required to polish a 
spring carefully made in this 

146. FoliBhin^ the Spring:. 
— Take a piece of wood or 
large pith, make it into ft 
cone at one ead, put a pin 
into the end to project half 
an inch ; put your spring - 
over this pin, and draw the 
outer end of the spring over 
tlie cone with the thumb of 
the left hand, as shown in 
Fig. 42, and with a well-worn 
brush, well charged with red 
stuff, you will, in a few 
minutes, polish the outer 
sides of the coils, turning the 
spring when one side ia ^- **~^o/^£ "'"*' ^^^^ 
polished, and repeating the 

process. The inner sides of the spring are more difficult 
to polish, and require care and practice. 

If the spring be placed on a piece of flat cork, and a 
finely-cut jwg be inserted through the centre, and pressed 
firmly on the cork, the spring will take the form of a 
cone (see Fig. 43), and by moving tlxe peg, with plenty of 



[Chap. XYI. 

Fig. 43.— To polish Inner Sides 
of Colls. 

red stuff on it, backwards and forwards, and also in a 
lateral direction, the inside of the spring will be polished 
in a few minutes. Care is necessary not to bend the 

spring, as if bent ever so 

little in the polishing it is 

useless. The flat edges of 

the spring are polished on 

writing-paper rubbed over 

with red stuffs, by pressing 

the end of the middle finger 

on the spring, and moving 

it gently in a circle. Cork 

or pith may be used for 

this purpose, instead of the 

finger. Wash the spring in benzole, and it is ready to 

blue. Care and experience are necessary to blue a spring 

well, but the practice will repay the time and trouble. 

147. Breguet Spring. — The tool for making the 
Breguet spring must be somewhat different, as only two 
of these springs are made at a time, to allow of the 
proper number of coils ; so all that is wanted is a slit 
cut across the pivot, or small part of the winder (Fig. 44). 
The pivot should 

project a quarter of (©) 
an inch through the 
box, and two, in- 
stead of three, small * 
holes should be 
made in the side 
of the box. The wire, to give the requisite number of 
turns in the flat spring, in a given size, must be flat and 
broad ; but when two only are made at a time, the wire 
should be very little flattened, or the coils of the spring 
will be too close together; although I do not recom- 
mend this, it being preferable and more usual to make 
three at a time, and to use the flattened wira 

148. Applying the Spring. — The spring can now 
be applied to the balance, without being pinned to the 


Fig. 44.— Winder for Breguet Spring. 


collet, but merely sprung on, as the centre should be 
small, and the time of the vibrations counted. It 
will then be seen if the spring is of the exact size 
and strength, and if so, it should be pinned in so 
that the inner and outer ends terminate in the same 
radius, forming a complete circle, or at "equal turns," 
as it is called. The vibrations of the spring are 
counted by taking the outer coil in a pair of tweezers, 
and holding it up until the lower pivot of the balance is 
just resting on the board paper, giving at the same time 
sufficient motion to the balance to make it vibrate the 
time required for counting ; but an experienced timer is 
able to tell almost at a glance by the distance the 
weight of the balance stretches the spring, when the 
outer coil is thus held up, whether it will do or not. 

149. Pinning on the Spring. — In all springing great 
care is necessary in pinning the spring on the collet. 
Always broach the collet hole before pinning in, 
and see that, in the case of a flat spring, the hole 
is parallel with the collet, and properly chamfered 
at both ends, that the pin has a flat side, and that the 
spring is not permanently pinned in till it is nearly true, 
for when once on the collet, much difficulty is expe- 
rienced in bending the centre inwards. Put the 
collet on an arbor, and in the callipers you will see if 
the spiral is quite true ; if not, bend it gradually until it 
is so. By putting the arbor in the turns the spring can be 
got flat. If it is required to make an overcoil or Breguet 
spring of this, after having ascertained by counting the 
beats that it is the right strength, proceed to turn up 
the overcoil as follows, and be careful not to get the 
spring out of flat or truth in the operation, as it requires 
great skill to get it either flat or true after the outer coil 
is turned over. Make a piece of brass wire into the form 
of the stud of a sprung-under watch, driving the other 
end into a small handle, and when you have drilled, 
broached, and chamfered the hole, take the outer coil of 
the spring through this hole and fasten it with a flat- 


sided pin opposite the hole in the collet. The outer coil 
can now be bent over the top any height it requires to 
be, and also inwards to suit the stud, without touching 
the other parts of the spring ; and with two or throe pairs 
of tweezers of different sizes, properly made with convex 
and concave faces, a little practice will enable you to 
make a very nice curve, when, if pinned in so as to 
terminate just where the centre springs from the collet, 
it will give the best results. Tie overcoil should be 
turned up not less than a full turn, being bent up- 
wards and over in a very flowing curve ; the sharp 
bends or angles frequently seen in springs of this form 
applied to Swiss Watches should be carefully avoided. 

150. Caution about Breguet Spring. — ^A Breguet 
spring should never be applied to a watch with an 
index. It is perhaps the best form of spring for a 
pocket watch, having all the properties in action of the 
cylindrical spring, and the great advantage of flatness in 
form, but any attempts at producing a good timekeeper 
with this spring and index pins will end in failure. And 
any attempt at getting time in positions by pressing the 
outer coil of the flat spring against the outer or inner 
pin is mere jobbing, and, even if successful, would require 
to be repeated every time the balance had to be taken 

For flat springs with indexes I would strongly recom- 
mend the plan of pinning a spiing into the collet, in 
order to get the stud hole and index pins to correspond. 

The end of the overcoil of a Breguet spring should 
run into the hole in the stud before being pinned in, and 
if the stud is screwed into the cock without the balance 
it will easily be seen if the jewel hole is in the centre of 
the hole in the spring collet, as it should be. This spring 
should also be pinned in at equal turns. 

151. Chronometer Spring. — To make a cylindrical 
spring for a marine chronometer : — Take a piece of round 
brass or German silver (the white metal stands the 
heat best), of about five-eighths of an inch in length; 


drill, and broach out a hole in it to about three-eighths 
of an inch in diameter, and turn it true on an arbor 
to about half an inch ; this is the usual size of an or- 
dinary " two-day " chronometer spring. Determine the 
number of turns you require and the space you wish 
them to occupy on the block, leaving it long enough 
to have a little plain space at each end after the 
spiral is cut ; send it to a fusee cutter with a piece of 
the wire you are going to use, and have a groove cut to 
fit the wire, not too tightly, and not deeper than half the 
thickness of it, to take thirteen turns. If on examina- 
tion this spiral be found perfectly true, drill two holes 
opposite each other, at the outside of the grooves, large 
enough to take a good strong screw, and tap them for a 
left-handed screw. Ascertain the length of wire the 
spring will take and cut it off the bobbin, as it will not 
in that way be so liable to injury from twisting, etc. 
The best way of winding the wire on to the block, is to 
put an arbor that will take your block into a mainspring 
winder (either a clock or chronometer spring winder will 
do). The block should be fixed on this arbor with a 
screw, as it is liable to slip round. Make one end of the 
wire fast to the block, and attach a good-sized hand-vice 
to the other. The weight of the vice will keep the wire 
straight and enable you to wind it on to the block flat 
and tight; when full, fasten the other end with the 
opposite screw. You here find the advantage of the left- 
handed screws, which draw the spring tighter, whereas 
right-handed ones would push it loose on the block. Use 
brass screws, as they do not become hard and so break in 
the holes. Some springers use brass tubing for their 
blocks, but this is bad, as you cannot broach it true. 
The block should never be thick, as, if so, it requires to 
be made so much hotter to harden the spring. It must 
now be covered with a thin piece of sheet copper or 
platinum, but any metal that will not melt will do, 
though platinum is best, as it lasts longer, and saves the 
trouble of making new covers so often. Holes must next 


be made in the covering for the two screw heads, and 
bound round with fine binding wire ; it will then be ready 
for hardening. Make a good coal fire in a grate, and, when 
it is well burnt, put a tin or iron box filled with charcoal 
on the top ; this is less expensive than filling the grate 
with charcoal. When red-hot, the charcoal will give 
out a slow, even heat, not too fierce, so that yoii will see 
exactly what you are doing. When the spring is at the 
required heat (a dull red), plunge it into cold water — cold^ 
and plenty of it On taking off the covering, the spring 
will be found to be a little loose on the block ; this loose- 
ness will not be prevented by using a steel block, as the 
wire cools much sooner than the block, and consequently 
retains its diameter at the instant of the spring cooling. 
Now undo one of the screws, draw the spring tight with 
your fingers and thumb, and make the screw fast again ; 
the end of the wire may break off, but the left-handed 
screw win hold the end of the spring tight. 

The spring should be tempered by blazing in oil 
while on the block, as before directed for watch springs, 
and allowed to cool ; it should then be perfectly true and 
white. The outer edges should be polished by putting it 
on a piece of pith or soft wood that will fill up the 
spring, and, while keeping the thumb on the lower 
end of it, brushing it upwards with a short-haired 
brush, charged with coarse red stuff. The inside and 
inner edges may be polished on a cork with a piece of 
wood and coarse red stuff, or fine emery, and the outside 
on the block on which it has been hardened on an arbor 
in the turns ; it should then be well cleaned in benzole, 
and brushed, when it will be ready for blueing. The 
greatest care and cleanliness are necessary in hardening 
and tempering a spring, and, above all, it should not be 
made too hot. I am satisfied from experience that a 
chronometer having such a spring will have little ten- 
dency to gain on its rate, as chronometers usually do 
when the springs are made too hot, however much they 
may be let down afterwards. A chronometer does not 


gain on its rate, as is generally believed, in proportion to 
the hardness of the spring, since, were this so, there is no 
reason why practical men should not have found the 
exact temper of a spring at which it would neither gain 
nor lose on its rate. It is known, however, that a spring 
not hardened and tempered loses on its rate, and one 
hardened and tempered usually gains, and hence the 

152. Blueing Springs. — There is no advantage 
gained by blueing a spring; it is not thereby kept 
free from rust.' Indeed it is said that steel, when 
blued, is in a state of incipient oxidation, and it is 
known that the blue spring is more frequently found 
rusted in ships' chronometers than the bright por- 
tions of the escapement ; but as it is customary to blue 
springs, I will explain how it is done. The best way 
to blue a cylindrical spring is upon a block kept for 
the purpose, and not used for hardening. The block 
should be solid so as to heat slowly, and the grooves be 
cut very shallow, and not fitting the spring too closely, 
in order that the air may have access to all parts of 
it. The spring should be fixed with screws, as if for har- 
dening, and the block be placed on end upon a blueing pan 
over a spirit lamp. If the parts of the spiing nearest 
the pan are colouring more than the upper part, the 
block must be turned upon the opposite end. Every part 
of the apparatus must be dry and hot before commencing 
to blue, and above all things it must be clean, for the 
least particle of oil or dust will prevent the spring from 
blueing evenly. Covering the block with a short length of 
glass tube will prevent external air currents from affect- 
ing the spring, and will keep the temperature uniform 
within the tube. When the spring is the required 
colour, set the block down to cool, and, if the spring has 
been well polished, it should be a very bright blue. 

153. Applying Spring to Chronometer. — The spring 
being finished, the most important part of the work 
has to be done, namely, applying it. If the ends of the 


spring that have been in contact with the block screws 
be nipped off, you can proceed to bend in the ends for 
the stud and collet. You must have pliers with concave 
and convex sides, having pieces of brass pinned or soldered 
inside them ; these pieces should be segments of a circle, 
and turned parallel, so that in bending the ends inwards 
they may be kept flat, and not bent upwards or down- 
wards, as pliers tiled this shape would certainly bend 
them. More tlian one pair should be used, with different 
curves or segments, so that the spring may be bent 
gradually and not more than is necessary. The hole in 
the spring collet (which is now seldom made by springers) 
should be about half the distance between the centre of 
the staff and the inner side of the spring. 

With the collet on a broach, the hole should be 
opened from the right-hand side, not parallel with the 
collet, but a little upwards, to suit the direction of the 
spiral. Pin the spring on temporarily, and if it is out of 
truth, unpin it as often as it is necessary to make any 
material alteration in the form of the curve ; when 
nearly true, make the pin a proper length, and pin the 
spring in tightly. 

The pins should be made with care, round, and as 
near the taper of the holes as possible, with about one 
third filed off the side, the flat side to go next to the 
spring, and the hole should be chamfered at both 
ends. By putting the collet (and spring) on an arbor 
in the turns, you will see where the spring is out. 
It wants careful bending ; always bend it inwards, 
until it is quite true; if it is bent too much, and the 
curve has to be straightened, the spring will be spoilt. 
The collet should now be put in its place on the staff, 
and the balance in its position in the chronometer 
frame. If the curve is of the proper shape (it should 
occupy three-fourths of a circle, and be at the end about 
, half-way between the centre of the staff and the outer 
circle of the spring), mark the stud where the spring 
touches it, and drill your hole there ; the spring will then 


nin into the hole, and if you broach the hole (as in the 
collet) at the angle of the spring, you will have it true 
and upright without any more bending, which should 
be avoided as much as possible. 

My reason for prescribing a length of thirteen turns 
is that such a spring, besides having a freer action, will 
be isochronous when pinned in at equal turns, while a 
shorter one — say of ten or eleven turns — will require to 
be pinned in about a quarter of a turn short of equal 
turns, the extra length, for the reasons given before, not 
requiring any diflference in the compensation. 

As there is no difference between the springs for a 
marine and for a pocket chronometer, except the size, the 
method of making the latter is precisely similar. 



154. Middle Temperature Error. — After the inven- 
tions of Earnshaw and his contemporaries had brought the 
ship's chronometer to great mechanical perfection, it was 
discovered that, though the ordinary compensation balance 
^Fig. 46) afforded an adjustment for a limited range of 
temperature so near that it was difficult to detect or record 
any difference, if the chronometer was subjected to a wide 
variation of heat or cold, a very serious deviation from 
its rate would be the result. The deviation was found to 
be constant, and was this, that if a chronometer be ad- 
justed for two given degrees of temperature, say at 30^ 
and 90° Fahrenheit, and is going to mean time at each 
of those degrees, it will gain two to three seconds a day 
at 60^. This method of adjusting the chronometer at 
two extremes of temperature and dividing the error in 


the balance, gave rise to some confusion of ideas, and 
the error was termed the " middle temperature error " 
by watchmakers ; most illogically, however, as the error 
is greatest in extremes. It was found that it arose from 
the compensation weights of the balance not moving 
towards the centre fast enough in heat ; and a theoiy 
was advanced, and was for a long time adhered to, 
that the balance spring lost its elasticity in heat in an 
increasing ratio. It was conjectured, however, that the 
error arose from some inherent incapacity of the balance, 
in the movements of its compensation weights to and 
from the centre, to accord exactly with the loss or gain 
in the elasticity of the spring in heat or cold. 

155. Berthoud's Experiments. — F. Berthoud found 
by experimenting on a chronometer with a plain balance, 
that it lost, in a range of temperature of 60" Fahrenheit, 
six minutes and a half in twenty-four hours ; but he does 
not appear to have made his experiments with a view to 
ascertain if the loss was equal for equal increments of 
heat, but to find its total amount, and to determine how 
much of it was due to the spring, and how much to the 
balance. He calculated the loss as follows : — 

Expansion of the balance . . . .62 seconds. 
Loss of the elastic force of the spring . .312 „ 
Elongation of the balance spring . . .19 „ 

Total 393 „ 

At a later period the late Mr. Charles Frodsham 
obtained a result nearly corresponding with this. Now, 
as the results show the loss from the inelasticity of the 
spring to be more than five times as much as that arising 
from the expansion of the balance itself, it is evident 
that the attention of chronometer makers should be kept 
directed to this point in their efforts to correct the error. 
It is obvious that if some material could be discovered 
applicable to the purpose, that would retain its elasticity 
for a sufficient length of time, and be less sensitive to 


changes .of temperature, both the primary and secondary 
errors would be greatly reduced. 

I have alreaidy stated the unadvisability of any ill- 
considered change of material for the springs of watches, 
but perhaps the palladium alloyed springs may at a 
future time come into use for marine chronometers, when 
the method of drawing the wire and setting them in 
shape of an equal hardness, etc., has been more fully 

156. Effect of Temperature upon Springs. — Taking 
it for granted that the elasticity of the spring varies in- 
versely with the temperature, and that the weights, or 
virtual radius of gyration, move in and out in the same 
proportion, it follows that the moment of inertia (which 
should follow inversely as the square of this radius) will 
not increase sufficiently fast in heat nor decrease suffi- 
ciently fast in cold. 

Sir E. Beckett shows this mathematically thus : — 

Let T be the distance of the compensation weights 
from the staff or axis of the balance, and let us call 
them both together m; for this purpose we have no- 
thing to do with the rest of the balance. Let dr be 
the increase of distance of the weights for some given 
decrease of heat. Then the new moment of inertia of 
the balance will be M (r^ + 2r dr -|- c/r^), and the ratio 

of the new inertia to the old will be 1 -f — "^ ( v) » ^^^ 

now the term ( ~j is too large to be disregarded as it 

may be in the similar formula for pendulums, because 
dr must be larger in proportion to r than it is in a 

Again, the ratio of the moment of inertia for an 
equal increase of heat to its amount of inertia in the 

2 dr ldr\^ 
middle state will be 1 "^ I ' ) assuming that equal 

successive increments of heat produce equal variations 
of r, which, however, is not quite the case, as it is in 
pendulums. Consequently, the increase of moment of 




inertia for a given rise of temperature is less than its 

(« ^ A 
— j , or the compeo- 

Fig. 45. 

sation fails to that extent in one of the three states of 
cold, middle, or hot temperature. 

Mr. Rigg, in a paper read before the members of the 
Society of Arts, Feb., 1879, says : — 

"It is not evident, without a little reflection, why 


the error is a gain at temperatures between those for 
which the adjustment has been made, and a loss at 
temperatures both above and below that range ; but the 
figure [Fig. 45] will at once show that such is the casa 
Assume the chronometer to be adjusted for 15° and 35^ 
Centigrade ; take two axes of co-ordinates, and let points 
on the vertical axis indicate temperatures ; through 
these points draw horizontal lines parallel to the other 
axis of co-ordinates. If distances are measured along 
the lines corresponding to 15" and 35", to indicate the 
tension of the balance spring, and a line be drawn through 
the points thus determined, the tension at any tempe- 
rature will be ascertained, on the assumption that it 
varies uniformly. The motion of the weights, however, 
and therefore the moment of inertia, does not vary 
uniformly, and must be expressed by points on a curve 
of some such form as that shown in the figure. Now, if 
the ratio of the tension to the moment of inertia were 
invariable, this latter would be determined for all 
temperatures by a straight line passing through the 
points on the curve at which it is made to correspond. 
The figure shows that, between these points, the tension 
is relatively in excess, causing a gain ; whereas, beyond 
them on either side, the converse is the case, and there is 
necessarily a loss. 

** The curve just discussed suggests a method of 
determining the manner in which the chronometer, as a 
whole, varies with the temperature. For assume it to be 
accurately adjusted at 15° and 35°, and then maintain 
it, for periods of twenty-four hours each, successively at 
a series of different temperatures ; the loss or gain due 
to the change will indicate the distance of the corre- 
sponding point on the curve to the right or left of the 
straight line. Representing now each second by, say, 
six inches on a large diagram, the observed rates may be 
plotted, and the curve obtained will at once indicate the 
gain or loss to be anticipated at any given temperature. 

^* And a further extension of this principle suggests 






itself. The abscissas of points on the line of tension 
represent forces, and the moment of inertia is also 
measured in the same terms. Hence, the interval be- 
tween these two lines corresponds to a force dependent 
on temperature, and, so long as it can be kept in a 
constant proportion to the tension, the rate will be 
invariable. Now, a variation measured in seconds really 
indicates a change in the proportion subsisting between 
the two forces, and a number of carefully-made obser- 
vations, through a long range of temperature, might 
enable the mathematician to formulate the law governing 
the motion of the weights, and thus to determine their 

157. Safe Range of Temperatura — Mr. Hartnup 

uses the following equation for obtaining the rate of 
chronometers under observation at the Bidston Obser- 
vatory :— ^^a-c{:t-t)\ 

\ where m is the rate of the chronometer at the observed 

i temperature ; T, the temperature where it has its greatest 

gaining rate ; <, the temperature at which it is observed ; 
a, the daily rate at the temperature T ; and c, the constant 
representing the change in the rate consequent on the 
variation of the temperature from T to T ± 1^. 

This method is a simplification of the former, the 
formula containing, as may be seen, only one variable 
term, involving the square of the difference of tempe- 
rature. In a letter to the Secretary of the Horological 
Institute, in reply to one from him on the subject, asking 
for particulars of the mode of procedure adopted at his 
observatory in rating chronometers, and published in the 
Horological Journal of June, 1878, Mr. Hartnup says, 
after a description of the method there adopted for 
maintaining them in constant given temperatures : — 

"The range of temperature to which we can safely 
expose a chronometer without the risk of sensibly 
changing the state of the oil appears to be so small, that 
we have limited this in our trials to the two extremes 



of 55? and 85^ (Fahr.), our object being to obtain the 
corrections due to imperfections in the balance apart 
from other sources of error. When limited to this small 
range of temperature, it is necessary to have recourse to 
scrupulous accuracy in obtaining the rates at the two 
extreme and middle temperatures, and the arrangements 
necessary for the accomplishment of this object are far 
more difficult to attain than those required for keeping 
uniform temperatures. 

158. Daily Bate of Chronometer. — " An example 
copied from our records will show that an error of two or 
three tenths of a second in determining the rate at the 
two extreme and middle temperatures, acting in opposite 
directions, would be quite sufficient to defeat our object. 

*' Example showing the Mean Daily Bate of a Chronometer on trial at this 
Observatory during a period of Twelve consecutive Weeks, 

No. 2720. 

Mean Temperature Fahrenheit. 

Week ending. 




1877, Oct. 20 . 

-f 2-478. 

, „ 27 . 


4- 2-438. 

, Nov. 3 

-f 2-17 

, „ 10 . 

-f 0-828. 



, „ 17 . 


+ 2-47 


, ,, 24 




, Dec. 1 


+ 1-97 


, „ 8 . 

+ 0-42 


, „ 15 . 


+ 2-28 


, „ 22 


+ 2-47 

, „ 29 


4- 1-82 


78, Jan. 5 

+ 0-47 




+ 0-57 

4- 2-20 

+ 2-59 

" It will be seen from the above that the rate of this 
chronometer in 70^ is not the same when the temperaturo 


is raised from 55^ to 70^ as it is when it is lowered 
from 85^ to 70^, the results being as follows : 

" Rate in 70^ when the temperature is changed — 

From 65« to 10^. From 86« to 70«. 
4- 2-478. 4- 2-178. 

4- 2-47 + 1-97 

4- 2-28 4- 1-82 

Means 4-2-41 4- 1-99 

*'We have the rate of this chronometer for the three 
temperatures, 55^, 70^*, and 85°. 

" To find the rate for any intermediate temperature — 

Let T = tlie temperature in which the chronometer has its 
maximum gaining rate. 

B = the rate at the temperature t. 

c = the factor, or constant number, which, multiplied by 
the square of any given number of degrees from t, 
shows the amount of loss for that number of degrees." 

Let us now apply the formula already given — 

m=i a — e {t — t)K 

Substituting the three rates and the three tempera- 
tures, we have 

•67 = R — (t — 56)« (i.) 

2 -20 = R — c (t — 70)3 (ii^) 

2-59 = R — c (t — 85)« (iii.) 

Now here are three equations from which can be 
found the three unknowns, R, c, T. 

Subtracting (ii.) from (i.) and (iii.) from (ii.) we have 

— 1-63 = — 2 c T (70 — 55) — (552 _ 702) /^^ j 

— -39 = — 2 c T (86 — 70) — c (702 _ 852) (v.) 

Again, subtracting (v.) from (iv.) and remarking that 

(70 — 65) = (85 — 70), 
we have 

— 1-24 = — c (652 __ 703 _ 702 ^ 852) -. _ 459 ^ 

.•• = -— =0002756. 


Again, by adding (iv.) and (v.) to find T, 

— 2-02 = — 2 c T (30) 4- (853 _ 552). 

4200 c 4- 202 ^ 

T = — -'- = 70 4- 12-2 = 82-2. 

60 c 

From (i.) a = '57 -f c (82-2 — 55)2 ^ .57 _|. 2-04 = 2-61. 

The following mode of solution, applicable to any 
example, will now be understood. 

Example : 

Eate inbb" =z-\- 0-57s. ... r, 

r — r' = — 1-63 ... rf. 

„ „ 70« = -f 2-20 ... /, 

/ — r" = — 0-39 ... ef. 

„ „ 86° = -f 2-59 ... r", 

rf— rf' = — 1-24. 

d-\- d' = — 202. 

^ ^ _2 (^-^) ^ -2:48 ^ ,.,^2756. 
302 900 

T - 70 = i±^' = — 2:?i_ = 12.2. 
60 X <J M6536 

T=70 4- 12-2 = 82-2. 

R = r' — (t — 70) li/ = 2-20 4- 12-2 X 00337 =2-61. 


T = 82-2. R=:2-61. c = 0002756. 

Let N = any number of degrees from t, then the rate at 

T±N = R — c X N^. 

Required the rate at 60<» : n = 22 2 and n^ = 492*84, there- 
fore the rate at 60*> = 2-61 — (0002756 x 492-84) 
= 1-25. 

The following are the rates obtained, as above de- 
scribed, for each five degrees, from 55° to 85** inclusive : 

Temp. 55<» 60*» 65<» 70«» 75* 80° 85o. 
Rate 4- 057 4- 1*25 4- 1-79 4- 2-20 4- 2-47 4- 2-60 4- 2-69. 

The maximum gaining rate of this chronometer is 



found to be at 82^, and the difference of rate between 
55° and 85° is 2*02 sees. If the maximum gaining rate 
had been placed at 70°, the difference between 65° and 
85° would have been only 0*62 sec. The factor c = 
002756, and this is about the average for a chronometer 
with the ordinary balance. 

159. Duration of Test. — The length of time neces- 
sary to test a chronometer efficiently for thermal adjust- 
ment depends on the quality of the instrument. In the 
example given the chronometer is a fairly good one, and 
live weeks would have supplied the data for making the 
calculations sufficiently accurate for most practical pur- 
poses. In a twelve weeks* test we have six changes to 
70°, three to 55°, and three to 85°. 

In a report to the Marine Committee of the Mersey 
Docks and Harbour Board, Mr. Hartnup shows that a 
chronometer going for a period of twenty-eight weeks 
has a total error of 334 seconds of time, whereas if it 
had been corrected for temperature in the foregoing 
manner, its error would have been 12 seconds only. 


" Columns 1 to 4 have been copied from the pub- 
lished rates of chronometers on trial at Greenwich in 
1878, for purchase by the Admiralty. The weekly sums 
of daily rates in column 4 are those of the chronometer 
lowest on the list in order of merit. The rates in 
column 5 have been calculated in the same way as those 
for No. la. Table III. [see pp. 234 — 51. Column 6 shows 
the difference between the calculated and the observed 
weekly sums of daily rates. The weekly range of 
temperature being large, it is probable that the mean of 
the maximum and minimum thermometers may not 
represent very accurately the true mean temperature, 
and an error of one degree in some of the low tempera- 
tures would cause a difference in the calculated weekly 
sums of daily rates of between two and three seconds of 








1 + 1 1 1 +!j; 1 ++ 1 1 + 1 





1 +++++++++++++ 
1 +++++++++++++ 
















++ 1 1 ++++++"1 1 1 + 




1 1 1 1 I++++I 1 1 1 



"'i'r°r""++++T"T 1 








74 =7.| =T."|"" 


" Difference between the first and second 14 weeks. 
Uncorrected for change of temperature = 334 seconds of time. 
Corrected for change of temperature = 12 seconds of time. 

" At the time of the trial the maximum gaining rate 
of this chronometer was +3-32 sees, in a temperature of 
87^, and tlie factor c was — 0*0045. The temperature 
used in the calculations is the mean of the maximum 
and minimum thermometers shown in columns 2 and 3. 
The weekly range of temperature is large, and a small 
error in the mean would cause a large difference in the 
calculated rate. For the week ending March 30th, two 
degrees of Fahrenheit would make a difference of five 
seconds in the calculated weekly sum of daily rates, the 
mean temperature for that week being up^wards of 40° 
from the temperature in which the chronometer had its 
maximum gaining rate." 

160. Airy's Experiments. — Sir G. B. Airy, in ex- 
periments which he made in the year 1859 with one 
of Charles Frodsham's chronometers, with a plain brass 
balance fitted experimentally for the purpose, found the 
loss consequent on each increment of heat to be 6*11 sees, 
for one degree (Fahr.) in twenty-four hours. Referrin<' 
to this result, Mr. Frodsham, in his report to H. M. Com- 
missioners of the International Exhibition, 18G2, says: — 

"The strength of springs being inversely as their 
length, and the effect on time as the square roots of 
their lengths, the loss amounts to about 17 seconds 
per diem for 60° of increased temperature. The decrease 
of rate by the expansion of the balance is also easily 
found by figures, it being inversely as the diameter, and 
in a plain brass balance amounts to about 63 seconds 
per diem for a change of temperature of 60° of Fahr. 
But 17 + 63 = 80 only, whilst the fact is established of 
a uniform rate of 6 "11 seconds for each degree of Fahr.; 
and since 6*11 x 60 = 367 seconds, it leaves 247 seconds 
of daily rate still to be accounted for, and this has been 
satisfactorily proved to arise from the spring's loss of 
elastic force by an increase of temperature." And again, 


after remarking upon the smallness of the error of 4 
seconds in a range of 60^, which the ordinary compen- 
sation balance fails (from its inherent defect) to remedy, 
and observing that if the motion were required to be 
greater than it is we should have a greatly increased error 
in the extremes, whilst the reverse effect would arise if the 
main error were less, Mr. Frodsham goes on to say : " The 
total decrease of the diameter of the compensation balance 
to restore the chronometer to time for a daily loss of 367 
seconds for 60*^ of temperature by a direct motion is 
but xo^oTT o^ ^*s diameter, or about yo%i7 ^* ^^^ compen- 
sation masses or weights. But it requires 80° of laminae 
to effect this object, because this motion is not direct, 
but by leverage, and moves diagonally, and consequently 
with loss of motion in the ratio of the hypothenuse to 
the perpendicular ; whilst, therefore, the direct decrease 
of the diameter of the balance of xAtt affects a com- 
pensation 6 min. 7 sees., the compensation weights or 
masses being moved nine times that quantity on the 
circular compensation rim would only correct the com- 
pensation to the amount of one second daily, supposing 
the weights to stand at 90° on the compensation laminae. 
Thus the travelling forward in heat and backward in 
cold, instead of towards the centre, is one of the chief 
causes of the error in the extremes. The forward or 
diagonal motion I have proved to amount to nearly the 
entire motion towards the centre of the balance. . . . 
These are the true causes of the error, and show that to 
obtain perfect and uniform compensation, the masses 
should move telescopically in the same ratio and uni- 
form manner that the plain balance is known to expand, 
and the balance spring to lengthen and lose its elastic 
force. '* 

161. Hartnup's Suggestions for Checking the Error. 
— Mr. Hartnup, in his report, says that after supplying 
the necessary data for making the calculations for varia- 
tions of the ordinary balance, where full advantage was 
taken of the information given, he found it necessary. 


in order to show the practical utility of his method, to 
furnish the mariner with the calculated rates for the 
various tern i)eratu res to which his ship was likely to be , 
exposed, and to supply him with the means of keeping 
a systematic record of the performance of his chrono- 
meters at sea. Following up this idea, he prepared 
a skeleton form for recording the observations at 
sea. The manager of the Pacific Steam Navigation 
Company had this form printed, and the ships of the 
Company were supplied with books containing these 
forms. For still further reducing the error, Mr. Hart- 
nup recommends that every vessel should carry three 
chronometers, the differences of Greenwich mean time 
between the first and second and the second and third 
being found daily, both by calculation and comparison ; 
any change in the rate of one chronometer will thus be 
shown by its agreement or disagreement with the other 
two. He says : — " The whole of this process is only the 
work of a few minutes, and a record of the calciilated 
Greenwich mean time, and the performance of the 
instruments relatively to each other, is preserved for 
future referenca" 

The following table shows that the loss that occurs 
beyond the maximum gaining rate of the chronometer 
increases as the square of the increase of temperature 
from that point multiplied by the factor c, which is 
obtained as before explained. 

For 10° = 

0-3 8 


„ 12-6« = 



„ 16« — 



„ 17-5° — 



„ 20« — 



„ SO*' — 



„ 40** = 



„ 60° — 



„ 60° = 



If the errors of chronometers were approximately 
tabulated as suggested, this would be the proper way to 


adjust them, and not by getting them to time at two ex* 
tremes ; but if a rate is given with a chronometer which 
it is supposed to keep, this plan of adjustment will not do. 
As Mr. Hartnup's calculations are based solely upon 
the ratio of the expansion and contraction of the 
balance to the elasticity of the spring, the other sources 
of error are not taken into account, and these somewhat 
modify in practice his conclusions, the error of nearly 
eleven seconds in a range of temperature of 60° being 
more than most chronometer makers have found in the 
ordinary balance. 

162. Adjustment of Chronometers. — Mr. Charles 

Frodsham says : — " With balances of the best proportions 
chronometers may be adjusted to within 4 seconds of 
daily error for a change of temperature of 60° Fahr., but 
if we adjust for extremes we gain in the middle tem- 
peratures, thus : — 

Thermometer 90° daily rate 0*0 seconds 
60° „ „ 4-2-6 gaining 

30° „ „ 00 

"If we adjust for heat and middle, we have the 
following : — 

Thermometer 90° daily rate 0-0 
60° „ „ 0-0 
„ 30° „ „ —40 losing 

" And if we adjust for the middle and low, we have 
the following : — 

lliermometer 30° „ „ O'O 
60° „ „ 00 
„ 90° „ „ 4-0 losing 

and most chronometer springers consider that the average 
error of the ordinary balance is from 4 to 6 seconds 
a day in a range of temperature of 60*^ Fahr." 

163. The Error of the Compensation Balance. — 
Sir K Beckett says, speaking of the error of the com- 
pensation balance : — "I believe it has never been dis- 



pated that old Mr. Dent was the first person to explain 
the cause of this error in the ^Nautical Magazine' in 

It is not probable that the men who first discovered 
the error of iJie Eamshaw balance knew as much about 
the cause of it as we do now, but long before Mr. Denfs 
explanation of it thej knew that if a chronometer 
was adjusted at any two given temperatures, and was 
going to mean time at those temperatureSy it would gain 
on this rate at any temperature between the two^ and 
lose on its rate if the temperature went beyond tiiem. 
And they knew that this arose from the fiust that the 
compensation weights of the balance were not carried to 
the centre fast enough in heat^ and receded from it too 
fast in cold. 



164. Hardy's Balanea— In 1804, WOliam Hardy, 
watchmaker, invented a balance (shown in "Fig. 47) 
wholly different in form to the circular balance in com- 
mon use. He called it a permanent compensation balance, 
and in his description of it he says : — " It will carry the 
weights to the centre quicker in heat tlian they are made 
to recede in cold." Hardy evidently knew what was 
wanted, although he did not succeed in obtaining it, as 
his balance never could have answered the description he 
gave of its action. However, he received the patocHUi^ 
of the Society of Arts, who gave him thirty guineas for 
his invention, and he afterwards obtained the gold medal 
of the Society and a prize of fifty guineas for the in- 
vention of an escapement which has been very little 
heard of sinca 

Hardj's balance (Fig. 47) consists of a diametrical 


liar, having two small pillars rising Tertically at its endn, 
upon which are screwed the compensation weights, a a. 
The bar is not a lamina, but conBists of two distinct 
pieces of metal, the pillars or standards being riveted 
to the lowermost, which is of brass, at » and « ; the top 
piece, which is steel, is £Ied thinner at the ends, the 
liveting of the standards holding it tightly against a 
shoulder which is left upon each for that purpose at the 

Fig. te.—Otilsuii} Compautlon BaJaiioe wltli BlidinE weiithts. 

ends. A boss is left upon and between these pieces at 
the centre, through which they are screwed to the col- 
let on the staff, in the same way as an ordinary 
chronometer balance, and these are the only places where 
the pieces are in contact. The steel bar is filed very thin 
at the ends, in order to offer less resiatajice to the brass 
when expiuiding, on the saine principle as having the 
steel thin in'tbe lamina of an ordinary balance. The 
compensation acts thus in heat : the steel bar not 
expanding so much aa the brass one, e e, the standards 
which carry the weights form a sort of lever, of which 
the ends of either liar may be considered the fulcrams, 
and the weights are drawn or driven to and from the 
centre or axis of the balance in heat and cold. But the 


Htandarda upon which thejr are fixed having no &ction 
in themBelveR, the weights cannot approach or recede 
from the centre in heat and cold in any other ratio than 
do those of an oi-dinary circular balance, to which this 
balance ia obvioualj inferior for various reasons. The 
screws, c c, are used as timing screws. 

This balance differed in form from any that had been 
made Ix^fure it, and su<;gested Bome others that were 

71g. v.— Hudj"! Biluio*. 

afterwards made on the same principle, which hare given 
the best results. 

165. Eiffe'B OT HoUnenz's Balance.— In ISiO Robert 
Molineux patented an auxiliary compensation balance^ 
but it happened that at the time he obtained hia patent^ 
E. J. Eiffe had a cbronomet«r on trial at Qreenwidi with 
auxiliary pieces exactly similar to those for whici he 
obtained his patent. For this, on the recommendation 
of the Astronomer Royal, Sir G. B. Airy, Mr. Siflb re- 
ceived from the Lords of the Admiralty a reward of jG300. 
As, however, the patent was secured by Molineux, this 
balance was generally called after him in the trade. 

In speaking of compensation balances some years ago, 
I incidentally mentioned this balance, describing it as 
Molineux's, and it was at that time, as before mentioned, 



F%. <&— Elffa's or Molmeni'i BaJuiog. 

that Mr. Eiffe, then an old man, wrote to me reminding 
me of the above grant, and of the deciaion of the 
Astronomer Itoyal in 
faia &vonr. 

This was the first 
auxiliary which secured 
practically the object 
in view, namely, the 
diminution of the in- 
ertia in heat, and vies 
versa, in the ratio of 
the spring's alteration of 
elasticity. It was veiy 
Bucceesful in the hands 
of ita first inventors, be- 
ing high up on the list 
at the Greenwich trials 
on several oocasiouB. 

Its action is shown in Figs. 48 and 48 a. Mr. Moli- 
neiix thus describes it : 
— "Having in the 
usual manner compen- 
sated the balance, so 
that its vibration shall 
be equal at the tempe- 
ratures of 30" and 
55° Fahrenheit, it will 
be found that if the 
temperature be raised 
to a greater beat such 
balance will vibrate, so 
that the cbronometer 
will lose its time or 
decrease its rate ; but 
by this invention I 
am enabled to com- 
pensate for the loss in the following manner : — 

" In Fig. 48 the balance, with its supplementary pieces, 

rig. 481.— EUCB'i or Holineni'i B»tauaa. 





is shown in the position it assumes at the temperature 
of 55^, its rim being then considered circular, and the 
middle projecting portions of the su^^lementaiy pieces 
banked in contact with the rim. Now, if the tempera- 
tare be raised, the balance rim, with the sapplementary 
oompensal^pieoes, wiU oomTmto the poidtion showL 
in Fig. 48a, in which, from the increase of heat^ the 
balance has ceased to be circular, its free endiE^ to 
which the adjusting screws are attached, having 
approached nearer the centre of the balance, and with 
them carried the free ends of the sapplementary com- 
pensating pieces, so that their middle projecting portions 
no longer bank on or are in contact with the inner side 
of the balance rim, and thus, by a proper adjnstment of 
the length, position, and weight of the sapplementaiy 
compensating pieces, I am enabled to compensate for 
temperatures above that at which the balance has been 
adjusted, while the adjustment for temperatures of 30^ 
and 50°, to which the balance had been before adjusted, 
remains imimpaired." 

Molineux made chronometers with this auxiliary, 
and one with a balance of this constniction was tried 
at Greenwich Observatory in 1840 for thirty 'weeks, in 
temperatures ranging from 18^ to 110^ Fahrenheit^ its 
error during that time being 11*4 seconds between the 
least and greatest, and 4*3 seconds between one week 
and the next. This was very good performance in such 
a long range of temperature, and with a constmction of 
the balance that is now thought but very little of; and 
it shows what may be done by careful manipulation, even 
where the principle is not to be commended Molineox's 
chronometers were very well and carefully made. I have 
seen them with a small ruby set in the auxiliary piece 
for the screw in the free end of the lamina to act 

* It may be of interest to notice that the fourth chronometer on 
the list at the annual Greenwich trial in 1884 was fumiflhad with it*i« 



These auxilmiy pieces act only in beat The; are 
made of two pieces of steel, which are screwed to the 
rim or to the arm of the balance, the outer half being 
left thick, and having holes drilled along it to allow 
of the screw which forms the auxiliary weight being 
shifted for adjustment, and a heel left projecting, which 
rests against the short end of the balance rim, and from 
that part backwards the pieces are filed to thin springs 
Assuming that the chronometer is adjusted and right at 
30" and 60" (without Uie auxiliary), il the temperature 
ie raised to 90" the rim of the balance will go nearer the 
centre, and the small screw in it will push the auxiliary 
piece, weight and all, before it, and, by diminishing the 
radius of gyration, cause the chronometer to gain the 
four BecondA it is calculated to lose with an ordinary 
balance so adjusted. 

166. Bent's Baluoe. — Although many modifications 
of this principle have been auda, aome of them with good 
results, it has the dis- 
advantage of acting only 
at high temperatnree, 
added to which, the 
mechanical action of 
parte in contact is unre- 
liable, the thin springs 
upon which the auxili- 
aries are carried being 
another weak point 
These considerations, and 
the necessity for great 
care and skUl in the 
operator to make this 
auxiliary of any use, 
render it unfit for ordinary work-a-day instruments. 
Dent took up Hardy's idea of a balanra, and made a 
|reat many varieties of it, of which the one shown in 
Fig. 49 gave the beet results. His object was to make 
a balance that would do what Hardy professed his 

tie. U.—DeBf UkUnw. 


capable of doing, namely, carry the compensation wei^ts 
to and from the centre quicker is heat and slower in 
cold, and also make ita action continuous: 

The principal compensation in this balance is in the 
diametrical bar, whi<^ is the osoal compound one of 
brass and steel, the brass being underneath t^e steeL But 
instead of the compensation weights being set on pillars 
rising from it, as in Hai'dj's balanoe, they are fixed to 

Vie. SD.— Hxtaap'm fiaUiiM, 

bent pieces of laminte, t t. Laving the brass inside; 
these laminae, upon the bending upwards of the bar in 
heat, move the weights more directljr towards the centrei 
The compensation may be adjusted by shifting the 
weights, V V, along the laminte, giving them more or 
less action, or by turning the laminB round a little on 
the bar for the same object The final adjustment for 
temperatures is made by screwing the weights to or 
from the lamiuK on their pillai-s. 

167. Hartnnp'e Balanoe. — Another modification of 
Hardy's principle is shown in the balance invented by 
Mr. Hartnup. Mr. Hartnup was not seized with a 
deeire of becoming famous in his generation as an iq- 


ventor in this way without knowing in what direction 
improvement was required, as is too often the case 
with inventors. But knowing wherein lay the weak- 
nesses of all the compensation balances that had been 
hitherto invented and made, he set about in a rational 
manner to avoid them, and suggested a new form of 
balance to Mr. Wm. Shepherd, a watchmaker of Liver- 
pool, which he thought would give more satisfactory re- 
sults. He laid down the rule that the balance must be 
circular, so that it could be made with facility in the 
lathe, and be more likely to keep its poise in the varying 
temperatures. After repeated trials, Mr. Shepherd 
produced the balance shown in Fig. 50. The centre 
bar, a a, is a lamina of brass and steel, the brass being on 
the top ', the parallel bars, b h, are also compound bars, 
with the steel in them uppermost, and these bars are 
solid with the rim, being fixed to the centre bar with 
four small screws on each side. The rim, d d, has the 
brass outside, as in the ordinary balance, but instead of 
being at a right angle to the bar, it is at an angle of 45 ^^ 
being what is termed " dished." 

The balance acts thus in heat: the centre arm 
bends downwards, while the two arms, 5 and 5, having 
the laminsd reversed, bend upwards, and the action of 
the two together is to throw the rim of the balance 
inwards all round, and thus diminish its circumference 
(and not in part only, as in the Eamshaw balance), and, 
from the form of the rim, the circumference diminishes 
more rapidly the more heat is given to it. The rim acts 
in the same way as the rim of the ordinary balance, but, 
from its form, in a less degree, and consequently, there is 
less of the error of the ordinary balance. 

In cold, the laminsB acting the reverse way tend to 
flatten the rim, and there is, therefore, less motion of 
the weights from the centre in cold than there is towards 
it in heat. The compensation weights, e e, are at the 
same angle as the rim, and the mean time screws, c c, 
are placed within the circle of the balance rim. It will 


ftt once be Been, therefore, that the action of Hub baluioe 
comes nearer to that of the plain nncut balance in ita 
expansion and contraction than any balance yet noticed. 
The adjuGtment for com[.enaation is efiected in the nsiul 
way, and although the movement of the w^^ta alimg 
the rim is less effective tLan in the ordinaiy balance, it 
affords a sufficient adjustments Several ^ronometen 
with these balances were made by Mr. Shepherd, whidi 
fulfilled the expectations of himself and Mr. Harbiap; 
the latter gave the rates of three chronometen with 
t^ese balances for a considerable period in temperatara 
ranging from 31° to 105°, which showed veiy litUe 
variation that could be attributed to imperfect otuit- 
pensation. But although the balance was given &eely to 
the trade, and a good many of thtan 'were made and 
applied at first, they did not seem to answer as well 
with others as Uiey had done with Mr. Shepherd, the 
original mt^er. The 
baUnoe vaa, for one 
thing, difficult to 
make, and oonse- 
qnently expensive— 
that is, in cximperison 
with the ordinary!)^ 
ance, — but the great- 
est drawback was the 
necessity of making it 
in two pieces, which 
added to the difficulty 
and uncertainty (rf 
turning it out perfect, 
eo that at present it 
may be said to be 
Fig. 61.— Pooie'i Auxiiiuy. abandoned, bat there 

is a principle in the 
bevelled rim which, I think, may yet be farther de- 

168. Pode'B Balance. — The balanoe known as Poole's 


auxiliary is, in fact, not an auxiliary at all, but is a 
hindrance rather than a help to the action of the balance. 
It was devised by the late Mr. John Poole, and consists 
of the application of a stop to the ordinary balance ; it is 
in its nature so simple that it needs very little describing 
(Fig. 51). The principle of stopping or interferiDg with 
the main compensation of the balance looks very un- 
promising at first sight, but it has had more success in 
practice than any other form of secondary compensation. 
The short segments of a circle of brass are screwed firmly 
on to the outside of the rim of the balance, opposite the 
arm ; the ends of the pieces that project towardbs the free 
end of the laminae are filed away from the inside, so that 
they are free of the rim, and one or two small screws 
are tapped into these pieces so as to be brought to bear 
against the rim of the balance at any given temperature ; 
they are the ordinary timing screws. Unlike most secon- 
dary compensations, this one acts only in cold ; the mode 
of adjustment is simple, and is just the reverse of 

If, in accordance with Mr. Frodsham's table, the 
chronometer is adjusted at 90^ and 60°, and is going to 
time at these two points, it will lose 4 seconds or more 
at 30°. If, then, the small screws in the check piece 
are screwed in until they just touch the rim of the 
balance at a temperature of 50^ or 60^, as the rim goes 
out with a lowering of the temperature these screws 
offer sufficient resistance to it, if in the right places, 
to prevent the loss of 4 or more seconds, which the 
balance would otherwise do. Mr. Poole made a great 
many chronometers for the trade, and with this balance 
a great many chronometer makers to the Admiralty. 
The Poole check has been successful in the hands of 
other chronometer makers, and from its simplicity and 
lowness of price has been more generally used than any 
other form of secondary compensation. 

169. LoBeby's Balance. — Mr. K F. Loseby obtained 
a patent for an improvement of a balance invented by 



TjC Roy, but Loseby's modification was so good in prin- 
ciple, and so exceedingly well executed, that it is the 
admiration of all chronometer makers. 

Le Roy's balance had small thermometers attached, 
the spirit which they contained expanding radially to 

the centre in heat. Loseby, 
however, substituted two 
curved tubes containing mer- 
cury sealed, with a little air 
in them, with the bulbs out- 
wards (Fig. 52). These bulbs 
were held in a cup-joint at 
the extremity of the ordinary 
laminated arm of the bal- 
ance, and by bending this 
joint inwards or outwards, 
Fig. 52.— LoBeby's Balance. the direction of the mercury 

in its expansion and con- 
traction could be altered and the adjustment completed. 
c c' are the laminated arms, to w' the compensation 
weights, 1 1' the timing screws or nuts, % i' the joints 
which hold the bulbs, and m m' the small mercurial 

The arms c c', by bending in in heat, alter ihe 
direction of the mercury in the tubes, which thus goes 
in more directly to the centre, and when going outwards 
in cold, have the contrary effect. 

Theoretically, this seems the best balance yet 
applied to a chronometer, and during the seven years 
Mr. Loseby sent chronometers with this balance to 
Greenwich, he held a very high place in the trials, 
notwithstanding that, as he was not a manufacturer on 
a large scale, he had few chronometers to choose from. 
He applied to Government several times for a reward for 
his invention, but his applications being unfavourably 
reported upon by the Astronomer Boyal, he never 
received one, and being ultimately disappointed and 
dissatisfied with what he considered unjust treatment, 



he, in 1853, retired from the Greenwich competitions 


170. Eollberg's Balance. — KuUberg's flat rim balance 

is another modification of the Hardy-Hartnup principle, 

as may be seen from the accompanying figure (Fig. 53). 

The arm and rim of this balance are in the same plane ; 

the bar is a compound one of brass and steel, the brass 

being uppermost ; the rim has the lamina reversed. 

The rim of this balance has much less action towards 

the centre than it would have if it were perpendicular, 

but it has a little, and the 

compensation is effected by 

the bar bending downwards 

in heat ; and as the rim is 

offering resistance and bend- 
ing upwards, the effect of 
the two actions is to tilt 

the weights, c c', which are 
placed at the extreme free 
end of the rim, upwards 
and inwards; the two 
screws, 1 1\ at right angles 

to the compensation weights, c c\ are the mean time screws 
or nuts. The balance is cut close to the bar ; the rim at 
its extremities has short circular segments standing up 
from it, which carry the compensation weights ; these short 
pieces give little room for the adjustment in heat and cold, 
but the compensation is also affected by the distance the 
weights are fixed from the plane of the balance, the two 
small screws seen in each weight being placed there to 
determine this height from the rim. Mr. KuUberg has 
been very successful with this balance in the Greenwich 
tiials, but it needs to be very well made, and there is 
considerable difficulty in fusing brass on to two sides of 
a piece of flat steel, as in this balance, soundly and without 
flaws. The bar and rim must not be thick, otherwise 
the compensation will not be active enough ; and as the 
compensation weights require to be at the extreme ends 

Pig. 53.— Ktdlberg's Plat-rim 


^ i 



of the rim, and the leverage from the balance staff to 
them is so great, the least external motion will give the 
weights a lateral vibration, and any damage they may 
receive from careless handling will alter the compen- 
sation. Thus, however perfect their performanoe may 
be when they are carefully handled, they are obviouslj 
less fit for the rather rough treatment to which chnmo* 

[ meters at sea are often subjected than a more rigid 

I balance would be. 

i Mr. Kullberg informs me that it is not with this 

balance he has achieved his latest successes, but with 

another, the particulars of 
^^^^^ ___. which he has not published. 

tfS^^SKHIiin^S 171. KnllbeqpB Impioyed 

Balance. — ^Another bidance by 
^ ^^^^ >j Mr. Kullbeig is shown in Fig. 

* ■ ^^k^— 1 54. he describes it as an im- 

proved ordinary balance, the 
auxiliary piece bein^ made from 
a part of the rim of the ordi- 
nary balance. Its action is 
k^^^ somewhat similar to that of 

Pig. 54.-ZnUb««^8 imptored Poole's balance, as it acts in 

cold only, and has a check- 
piece ; but unlike Poole's check, which acts on the entire 
balance rim, this one checks only the secondary compen- 
sation pieces, leaving the main adjustment free. I have 
seen no record of the performance of chronometers with 
this balance, but it seems to have all the elements of a 
good one, namely, simplicity and strength. 

The main compensation weights and the mean lime 
nuts act in no wise differently from those of an ordinary 
balance. The rim is cut nearer the middle than usual, 
giving about two-thirds for the main compensation, 
and one-third for the auxiliary ; the short section of l^e 
rim is divided into two by a slit cut in the middle, from 
the mean time nut to the screw A, which acts as the 
auxiliary weight. When the balance is being crossed 


out, a portion of the steel from which the arms are 
formed is left projecting to strengthen the bottom half of 
this short piece of the rim, at the extreme end of which 
a knee is left at c, and a small screw is tapped into the 
bottom piece, which is cut through between the screws A 
and B ; the top piece with the knee attached, and the 
screw A, form the auxiliary, the action of which is easily 
seen from the diagram. As the bottom piece is rigid, 
the top piece, or auxiliary, is prevented firom going out 
in cold by the knee o coming in contact with the point 
of the screw b. If, therefore, the chronometer be adjusted 
by the main compensation weights at 50^ and 90^ (which 
may be done before the short section of the rim is divided 
into two), when the auxiliary is separated irom. the 
bottom piece, if the screw b be brought close to the knee 
c at a temperature of dO"", it will prevent the auxiliary 
from going out farther if the temperature is lowered. 

The lower part of this short piece might have some 
action, according to the thickness of the steel attached to 
it, and need not therefore be a rigid stop, as if it had any 
action, it would be in the same direction as the other por- 
tions of the laminse, and the means of correctly adjusting 
the secondary compensation will depend on the resistance 
offered by the stop and the weight of the screw in the 
auxiliary pieca 

As the auxiliary piece is too narrow to have more 
than one hole drilled for the screw without weakening 
it, a weight, made to slide on it like the main compen- 
sation weight, would facilitate the means of adjusting the 
secondary compensation. 

172. Mercer's Balance. — Mr. Mercer modified the 
Molineux-Eiffe balance, and obtained very good results 
from it through a long range of temperatura As shown 
in Fig. 55, instead of the auxiliary weight being attached 
to a thin spring, to be pushed inwards by the free end of 
the rim of the balance, or, as in some of Molineux's 
balances, to a short piece of lamina screwed to the rim of 
the balance, he had a piece of lamina made, three-quarters 



Fig. 55.— Mercer*8 Balanoe. 

of an inch long, about the substance of the rim of a 12-size 
watch balance, fastened to which is a small sole-piece, 
which is screwed to the arm of the balance. These laminse , 
act in heat, and are stopped in cold hy the screws h h\ 
which can be adjusted so as to stop them at any required 
temperature. The screws m m' are the secondary com- 

pensation weights, and this 
compensation can be made 
more or less efiectiTe by 
shifting them to or from the 
ends of the laminae, in which 
are a series of small holes 
to receive them. This aux- 
iliary acting only in heat^ 
and not being constant^ the 
mode of adjusting it is the 
same as that of most others 
of a similar principle. If 
the main compensation from 
the coldest point determined on to 50® or 60® is effected 
with the auxiliary pieces taken off, or by screwing the 
small screws A A' in until they prevent them from acting, 
the secondary compensation is adjusted by setting the 
screws h h' in such a position that the auxiliary lamins 
cannot go out in temperatures beyond the highest to 
which the main compensation has been made, but they 
are free to go in beyond that temperature. 

Mr. Mercer has been very successful at the Greenwich 
trials with this balance, and it may be credited with 
making several of the numerous chronometer makers 
to the Admiralty. 

173. Ally's Compensation Bar. — For a long time 
Sir George Airy advocated some means of acting on 
the balance spring in order to enable a final adjustment 
for compensation to be made without the necessity of re- 
moving the balance from the chronometer, and, some years 
ago, he invented what he called a compensation bar, which 
at first he required to be put to all chronometers sent to 


the observatory for trial, but which condition he after- 
wards modified. This arrangement consists of a steel 
radial bar fixed friction-tight on the balance staff under- 
neath the spring collet. The extremities of the bar have 
short springs fixed to them, and carry two small weights ; 
they are set on so as to keep these weights slightly pressed 
against the inside of the rim of the balance. By moving 
the bar round on the staff, these weights are brought 
nearer to or farther from the free end of the rim, and 
affect the compensation somewhat after the manner of 
Molineux's invention. 

The trade disapproved of this, however, and it has 
never been employed except in chronometers sent to the 
Greenwich trials. 

174. Chronometer Balances. — Chronometer balances 
may be divided into two classes, the vertical and hori- 
zontal. The former, including all the balances made after 
Hardy's idea, are certainly the more perfect theoretically, 
as they admit of constructions that comply with the con- 
dition of the weights going to the centre in the necessary 
ratio, but they are all more or less weak ; and if the rim 
is bent out of its original shape, either by an excess of 
heat or cold, or by accident, the compensation will be 
altered ; they are also more difficult to make, and have 
never come into general use in any form. The horizontal 
balance has a known, and so far an incurable, defect in 
principle, and all auxiliaries applied to it are discon- 
tinuous, acting either in heat or cold only ; but in other 
respects it is so well adapted to a portable timekeeper 
that it is likely to hold its ground, with all its defects. 

175. Chronometer Trials. — The annual trials of 
chronometers at the Royal Observatory for purchase by 
the Admiralty give an undoubted stimulus to chronometer 
makers to improve the science of horology, although with 
the yearly increasing number of steam-ships, and conse- 
quently shorter voyages, chronometers are of less import- 
ance to navigators than when voyages extended over 
longer periods. It would, however, be a great improvement 



on tlio present system if no cbronometer were to be received 
for trial except from and in the name of the band fde 
maker of tlie instrument, and if the Liords of the Admi- 
ralty were to offer, as they formerly did, rewards of three, 
two, and one hundred pounds for the first three chrono- 
meters on the list, instead of taking it for granted that bo 
further improvement is possible, and paying such a price 
for the best as makes it worth the while only of those 
who desire to have the title of chronometer maker to 
the Admiralty to com^>ete. 



176. Caron's Keyless Watch. — The first keyless work 

of which we have any account was made by Peter Caron, 
a celebrated Paris watchmaker (afterwards styled Beau- 
marchais), for Madame de Pompadour about the year 
1753. It is thus described by him : — "A watch in a 
ring ... It is only four lignes in diameter, and 
two-thirds of a ligne in height between the plates. To 
render this ring more commodious, I have contrived, in- 
stead of a key, a circle round the dial, carrying a Jittle 
projecting hook. By drawing this hook with the nail 
two-thirds round the dial the ring is rewound, and it goes 
for thirty hours.'' 

177. Other Contrivances. — Many contrivances have 
been imagined since this invention for dispensing with 
the key in winding the watch, such as by drawing out the 
push piece in the pendant, and pushing it back again, and 
opening and shutting the cover of the case, (fee. ; but they 
were all more or less complicated, took up a great deal 
of room in an instrument, where there is never much to 
spare, and were consequently out of favour, especially 

Chap. SIX.) 



with watchmakers, and never generally made. They were, 
in fact, merely mechanical toys, constructed from time to 
time for amateurs, or for persons of distinction, more as 
a proof of the maker's mechanical ingenuity than for 

Fig. 56.— Prest'B KejIsnWaA. 

anything else. The Emperor Napoleon poBaessed one of 
these watches, which is said to have been wound by the 
motion given to a pendulous weight in walking, as in a 

178. Preat'fl Keyleas Wort— In 1820 Thomas Prest, 
of Chigwell, in Essex, took out a patent for " a new and 
additional movement applied to a watch to enable it to 
be wound up by the pendant knob, without any detached 
key or winder." Fig. 56 represents Prest's winding ar- 
rangement, of which the following (taken from his specifi- 
cation) is a description ; — 


I . 

:. I 



Fig. 56, 1., represents the back of the pillar plate wbei 
the dial is taken off. Fig. 56, II., represents the actioi 
of the pendant pinion 3 into the under part of th( 
wheel B, which is not so distinctly seen in Fig. 56, 1. 

In Fig. 56, 1., A is a brass wheel, fixed on the square oi 
the barrel arbor, and secured by the pin, 2 ; b is the wheel oi 
communication between the pendant pinion 3 and wheel a 
This wheel b has the same number of teeth as wheel A 
but the underside next to the pillar plate is hollowed, sc 
that the teeth have a double action sideways, or laterally, 
into the wheel A, and downwards, or vertically, into the 
pinion 3 (for this vertical action see Fig. 56, IL , where the 
teeth of wheel b act above the pinion 3). 3, is the pendam 
pinion, whose arbor 4 passes through the pendant shanl 
8 ; the end of the arbor is squared to receive the knol 
5 fitted thereon, and secured by a nut. In Fig. 56, II. 
7 is the arbor of the wheel b ; 5, pendant knob, whicl 
being turned gives motion to the wheel b, which tumi 
wheel A, which turns the barrel arbor, and therebj 
winds up the mainspring. 6 and 6 are two clicks whicl 
secure the wheel A, and consequently the barrel arbor ii 
its place, till the mainspring developing itself requires 
to be wound up again, by turning the pendant knob as 

I Brest's iDvention was to some extent adopted by the 

trade at the time, a good many of these watches being 
made for or by J. R. Arnold, and it is the undoubted 
parent of the present keyless watch ; but great improve- 
ments having been made in the lever escapement by 
Savage and others, and a demand having arisen for more 
accurate timekeepers, keyless watches were abandoned, 
as Prest's mechanism could be applied only to going 
barrel watches, and for many years (embi^acing the time 
of the greatest prosperity of the watch trade in England) 
no keyless watches were made here. 

179. Swiss Keyless Work. — The Swiss, however, see 
ing the applicability of the keyless mechanism to going 
barrel watches, had, about the year 1851, contrived many 

I : 

. i 


Fig. 57.~Bwln Kej\tea Work. 


r I 


I , 


methods of winding the watch and setting the hands from 
the button, for which Prest's winding arrangement had 
no provision. Fig. 57 shows the arrangement generally 
adopted by the Swiss in their best watches, but many other 
forms of keyless mechanism are applied by them to cheap 
watches. In the diagram the winding wheel and the wheel 
on the barrel arbor square are removed from the plate ; 
the winding wheel has two sets of teeth, one set gear with 
a small intermediate wheel which gears with the barrel 
arbor wheel, and the other which are contrate, with the 
winding pinion. The winding pinion r r' occupies a part 
of the round part of the winding stem a at «', upon which 
it can turn, but is prevented from moving up or down it 
by the shape of the plate, which has a part cut out of it 
to admit the pinion ; the lower part of it, r', has contrate 

f \ ratchet teeth cut upon it. The pinion r" v is fitted on 

I the square of the stem A, and its ratchet teeth are pressed 

upwards into those of the winding pinion by the spring 
m, as shown in b. By turning the button to the left (in 
the figure) the pinion r r" is rotated and the watch is 
wound, and by turning it the reverse way a ratchet action 
is produced between the ratchet teeth of the two pinions 
which click over one anotheir. In setting the hands, the 
thumb nail is pressed upon the push piece p which pushes 
in the spring m, and the pinion r" v down the square of 
the stem gearing its contrate teeth v with the intermediate 
wheel I (which is in gear with the minute wheel q), as 
shown in c ; when the stem is turned, the hands are thus 
set, the pinion r r" being out of action. The screw h 
prevents the long spring from rising from the plate, and 
the screw or stud n prevents it from pressing the pinion 

I t" V too deeply into the wheel Z ; i is a spring brace 

screwed on the other side of the plate lying across the 

' Blot in the stem s ; by withdrawing the screw at the end x 

of this brace a little the winding stem can be drawn out 

from the watch. The stem is fitted easy in the pendant 

and has a pivot at the other end which goes into a hole 

• in the plate or in a stud underneath the wheel I ; the 


Chap. XrX.] KEYLESS WORK. 263 

push piece is fitted to a pipe soldered into the case and 
having a part of it projecting in order to protect the push 
piece and prevent it from being accidentally pressed in 
in wear, etc. ; it has a V notch filed across it to allow of 
the nail going down. The objections to this form of 
keyless work are that it is not strong, is troublesome to 
make, and the contrate depths are rough, especially that 
of the winding wheel and pinion, which wear quickly as 
the pinion is always pressing the wheel away from it. 
These faults in the Swiss keyless work being apparent to 
English watchmakers, not many of this description have 
been made here, but what is known as the "rocking bar" 
mechanism has been invented and adopted as the English 
model. At first making the keyless work was a branch 
of watchmaking proper, and, the movements and calibers 
being different, great diversity of form and planting was 
the result. To Mr. Hewitt, of Prescot, the credit is due 
of having put a stop to this, as he made and planted the 
steel wheels on the movements. Other movement makers 
followed his example, and now almost all the keyless work 
is done in Lancashire by the movement makers, thus 
ensuring the best model, greater uniformity, and a great 
reduction in the price of the work. 

180. Rocking Bar Keyless Work. — Fig. 58 shows the 
rocking-bar keyless work, o is the bevelled conical wind- 
ing pinion ; 6 is a boss which is screwed to the plate, 
fitting through the countersunk winding wheel w, and a 
steel bar or platform c, upon which are screwed two 
wheels, i, h, which rotate upon two pipes projecting from 
the bar, the heads of the screws being countersunk ; the 
rocking bar c rests on the plate and the winding wheel w 
rotates on the bar ; the spring 8 on the other side of tihe 
plate keeps the winding in action by pressing a stud 
screwed into the bar and projecting through a space cut 
out of the plate. On turning the button to the left the 
winding wheel w is turned and the watch wound; on 
reversing the button the barrel-wheel is prevented from 
being affected by the click m, and a ratchet action 


takes place betveen t^e toetJi of the wheels i and g. 

T<g, SS.—'BocklBg Bar Karlen Wotk. 

The hands are set by pressing in the posh piece p 
with the nail, which throws the wheel i out of action 

Chap. XIX.] KEYLESS WORK. 265 

with the wheel ^, and the wheel h into action with the 
intermediate wheel I, which is in gear with the motion 
wheels, when by rotating the button the hands can be 
set either backward or forward. The intermediate 
wheel I is planted on the circumference of a circle drawn 
from the centre of h, with a radius to the centre of h ; 
this is to enable the hands to be set correctly, and to 
prevent them from being shifted when the wheel moves out 
of action from the other (as would be the case were it to 
move in and out in an oblique direction) ; the additional 
play of the extra wheel also furthers this object. The 
depths of the wheels i and h with g and I are adjusted 
by means of an oblong slot cut in the bar c which re- 
ceives the head of a small screw fixed in the plate; 
neither of these depths should be pitched deep; for 
setting the hands it is not necessary, and that with the 
wheels i and g has a tendency to become deeper, the 
action of winding pressing the end of the bar towards g. 
The intermediate wheel h should be just free of the teeth 
of the small wheel h when the winding button is turned 
the reverse way ; if it is at a greater distance from it 
the push piece p will be too long. The spring 8 must 
be strong enough to keep the push piece in its place. 
The winding pinion o has no fitting in the movement 
like the stem to the Swiss keyless work, but is fitted in 
the pendant of the case; it may have the teeth cut 
straight on its face or on a bevel ; either will make a 
smooth action if the pinion is the proper size, but the 
bevelled teeth are better and stronger. (See page 62.) 

If the watch is to have a hunting case where the 
cover is released by pushing in the winding button, 
the pinion and the stem on which the button is fixed 
must be in separate pieces, and, as it is not possible to 
get a large hole in the pendant, it is necessaiy to be very 
careful in arranging the sizes of the stem and the pinion 
arbor to keep them sufficiently strong. If the hole in 
the pinion arbor, or pipe, is left round, two oblong slots 
are cut in it opposite to each other, and a pin through 


; { 
• .1 




the stem projects into them flush with the oatside of the 

j)il)e ; this piu carries the pinion round when the button 

is moved, and at the same time allows the stem sufficient 

action in the pinion to press in the case spring and 

release the cover which the spring holds down. The 

I play allowed or the length of the slot should be as Uttle 

j , as possible, as the winding buttons are the reverse of 

"i ornamental, and should be worked as close in to the 

!i middle of the pendant as possible, as if left near the 

* middle of the bow, as they often aj-e, they are 

awkward to wind and most unsightly. This is a very 

good Wiiy of fitting a pinion, if done carefully. The pipe 

;" of the pinion must not be left too thin ; the hole in the 

stem must be broached and the pin carefully fitted and 

hardened and tempered, care being taken that it does'not 

j i project at the ends beyond the outside of the pipe. 

■ ■ The winding pinion is also often made 'with a square 

hole in it, and the lower part of the stem fitted to this 

square; when done in this way, the pinion face is 

hollowed out, the pinion being kept in its place by a 

small nut screwed on to the end of the stem and fitting 

the hollow in the pinion face ; sufficient play must be 

allowed for pushing in the case spring and releasing the 

cover. The winding button is let on to the upper part 

of the stem, which is squared to receive it, and kept in its 

place by a nut or a screw, countersunk, but the nut is much 

better and stronger, and a screw should never be used 

There is a cheaper way of making the rocking bar 
keyless work by putting the bar on the top of the wheels, 
as there is then no necessity for pipes to carry the two 
small wheels nor screws to keep them in their places, the 
wheels being merely fitted to pins screwed into the bar, 
and lying on the plate when the face of the watch is 
upwards ; but this method, although often adopted, is a 
very bad one, since it brings the winding wheel so near 
the centre of the pendant that only a very small winding 
pinion is possible, the saving of labour is not great and 
certainly not worth t\i© sa^xVicLQi^ ol y^x^ssv^^ Vs^^^lved 

Chap. XIX.] KEYLESS WORK. 267 

181. Fusee Keyless Work. — The demand for keyless 
watches, which has arisen within the last few years, has 
greatly reduced the number of fusee watches made in 
England, on account of the greater trouble involved in 
applying the keyless mechanism to them than to those 
with the going barrel The reason for this is that whHst 
the arbor of the going barrel remains stationary, except 
during the time of winding, the fusee and its arbor are 
continually turning one way in the going of the watch 
and the reverse in the winding; the keyless wheels must, 
therefore, be detached from the fusee when it is not being 
wound, which necessitates a somewhat complicated wind- 
ing mechanism. Many very ingenious contrivances for 
this purpose have been devised ; but the majority of 
them, owing to their being complicated and liable to get 
out of order, have served only to make the fusee keyless 
watch impopular. No winding arrangement for this 
watch is of the least use that is not strong and simple, 
and I only know of two or three that comply with these 

Fig. 59 is a diagram of a fusee keyless arrangement, 
which I have made and tested for many years, and found 
to answer admirably ; the merit of its invention has been 
claimed by several people. It is stronger than any other 
I know of, and has the great advantage of having the 
winding wheel and the wheel on the fusee arbor always 
in gear with one another, which obviates the risk 
of the teeth butting and stripping on these wheels 
coming into action (a fruitful source of failure in keyless 
watches with a shifting wheel, such as in the rocking 
bar mechanism, when improperly made). It is not 
difficult to make, the only part of it requiring great 
accuracy being the spring x x x" in the wheel on the 
fusee arbor. The winding wheel w turns on the stud s, 
which is screwed to the plate through the rocking bar, as 
in the previous arrangement. The detachment of the 
fusee while the watch is going is effected by the compound 
wheel R, shown in detail in Fig. 59, a, B,G^«j2Ldvcv.^^-^i^aia^ssto. 


Tig. 59.— rnaee Keylt 

Chap. XIX.] KEYLESS WORK, 269 

in D. When the push piece is released, the spring , presses 
the end of the rocking bar into the ratchet wheel r, and 
prevents it from turning when the watch is being wound, 
acting as a click when the button is turned the reverse 
way; the point of the bar is properly shaped for this 
purpose, and should not extend beyond the point where 
a tangent, drawn from the centre of the winding wheel, 
touches the circumference of the ratchet wheel, or it will 
be liable to be driven out during the winding. The 
ratchet wheel, r (Fig. 59, b), has a sink turned in it, into 
which a spring, x x' x'\ is fitted ; this spring ha« a boss, 
t by rising out of its plane. The wheel m (Fig. 59, a), 
has a large hole turned out of its centre, and fits over 
the central ridge, a, which is left standing up from the 
middle of the ratchet wheel r, upon which it can turn, 
and a triangular space, i, cut in it, into which the boss 
on the spring projects flush with the countersink, and a 
further projection, t, forms a tooth and comes flush with 
the upper side of the wheel m. The ratchet wheel, y, 
(Fig. 69, c) fits into the countersink just free of its edge, 
and screws on to the end of the fusee arbor, on which a 
right-handed thread is cut, and prevents the other wheel 
from rising from the plate. Upon turning the button in 
winding, the wheel m (Fig. 59, a), is turned to the right ; 
the point of the rocking bar prevents the ratchet wheel r 
from turning, and the inclined plane of the triangular 
space, iy forces the boss, which projects into it, down it 
until the projecting tooth, t, slides into a space of the 
ratchet toothed wheel /(Fig. 59, c), which it then forces 
round and turns the fusee ; when the button is released, 
the spring, x a;' x", returns to its former position, 
and the fusee is completely detached from the wind- 
ing work, the wheel / turning with it on its arbor 
without affecting the other wheels. The spring, x x' 
x", must be of such a ' shape between the points, x' 
and x"y that it will be strong enough to keep it fric- 
tion tight in the wheel r, and it must be weak enough 
between the points x and x to allow of the tooth. 


tf being drawn into the wheel, y, before the spring is 
drawn round in the winding; it should also be strong 
enough to prevent any slight action of the button in 
wear from driving in it, or it will interfere seriously 
with the good performance of the watch ; but it must not 
be too strong, or it will make the winding hard. By 
adopting the arrangement shown in Fig. 59 of having a 
small intermediate wheel on a stud screwed into the 
plate, and the minute wheel on the other side of the 
cannon pinion, more room is secured, and a proper-sized 
minute wheel can be fitted without interfering with the 
other parts of the watch. 

182. KuUberg's Fusee Keyless Work. — Fig. 60 

shows an arrangement of fusee winding by Mr. Kullberg, 
the drawing and description of which are taken from the 
Horological Journal of April, 1869 : — 

A is the fusee wheel, b is the winding wheel, c is the 
setting wheel, D is the platform, E is the stud for holding 
\.\ the platform and winding wheel with perfect freedom 

,. J I against the plate, F is the spring for holding both the 

I j winding and setting wheels out of gear, g is the push 

pin for setting the hands, H is the elastic screw for 
preventing the wheels on the platform from revolving 
until the winding is in gear with the fusee wheeL By 
turning the winding pinion (which runs through the 
pendant) to the right (to the left in the figure) the 
winding wheel moves slightly round the setting wheel c, 
which is kept stationary through being pushed against 
the rounded head of the screw h. "When the winding 
wheel is in gear with the fusee wheel A, the pin i, at the 
end of the platform, touches the outer end of the notch 
K ; and by a little further moving of the platform on the 
fulcrum I, the setting wheel is lifted free of the screw H, 
and revolves without any impediment whatever. For 
setting the hands, the platform guided by the pin i in 
the notch k, is pushed in so as to make proper depth 
between the wheels c and m. At this position the 
platfoim touches the pin L, thereby preventing the 


winding wheel from touching the fusee-wheel in the act 
of Betting. Locking the keyless work out of action, 

Fig. BO.— KuUbanCB Fnaee Eejlesa Wort. 

if sach is required, is accomplished by causing the cover 
or back of the case to pu^ in the push pin ahghtly, 
thereby releasing the setting vheel from die screw n ; 




the winding pinion will then turn round freely without 
driving into action — but since the setting 'wheel can by 
no means revolve until the winding w^heel is in gear wiiii 
the fusee wheel, the out-driving spring can be made tG 
act with nearly all its power on the winding end of the 

I platform, and consequently drive out with great force ] 

by turning the winding pinion or button to the left, the 
setting wheel is lifted, so to say, from the screw end H, 
and passes with a gentle click. If the push pin foi 

; { setting the hands is required to lock in during setting, 

I ^ a thin spring is fixed inside the band of the case, whicli 

slips up between the push pin and the case, and keeps 

i J the wheels in gear without pressure of the finger. A 

fine upright pin at the end of the lock spring causes the 
cover to push down the spring, and unlocks the pusl 

183. Chalfont's Fusee Keyless Work. — The above 

is a very good keyless work, but is rather difficull 
to make properly. Another arrangement, invented bj 
Mr. W. W. Chalfont, is shown in Fig. 61 ; it is the 
same in principle as the foregoing, and is thus de 
I scribed by him. "The wheel No. 3 always advances 

in the same position ; the pin No. 5 is so adjusted thai 
the wheel No. 3 never presents more than one point at 
the moment of touching. No. 2 is a platform with 
' a slot (No. 1) cut in the centre, to allow it to move 

from right to left. A screw regulates the distance so 
moved ; the head of this screw holds the platform down, 
allowing it the right freedom. No. 3 is a wheel running 
on a boss under the platform through which the slot is 
cut This wheel is in gear with the contrate wheel 
^ No. 8, and with the pinion, which is also run on the 
platform. No. 11 is a spring which beans upon pin 
No. 14, and throws the platform to the right, and 
the wheel No. 3 free of the wheel No. 4, which 
latter is fastened on the fusee, at the same time 
pressing the pinion into gear with the pin No. 5. H 
the button attached ^ t\i<B wsuXx^^fe ^W^l No. 8 is 


turned to the right (left in the diagrani) to wind, the 
pinion cannot turn round because thd pin No. 6 is 
between the teeth, therefore the platform is forced across 

Fig. ei.—Clialfcmt's Fneee Keylesa Work. 

from right to left, and the wheel Na 3 into gear with 
the fiisee wheel Ho. 4. By this time the pinion is 
drawn free of the pin No. 5, by means of one end of the 
platform coming agiiinst pin No. 10, leaving it quite free 
to wind Aa the button is released it is thrown back 
again. In turning the button the reverBe way the 
pinion trips over the pin No. 6. For setting hands, 
when the piece No. 12 is pushed in, it mov«& ti« 


pUtform No. 2, and carries the pinion free <^ the pi 
No. 5, and straight into the minute wheel No. 6, at tl 
same time bringing the end of the platform romid tt 
pin No, 7, locking it free of the fusee. Now the haoc 
can be set For shutting off this winder, let the ca! 
cover move the stud Na 9 on the push piece snffidentl 
in to free tlie pinion of pin No. 5," 

184. The Various Forms of Eeyleas Work.— Thei 
are many other forma of fusee keyless work which : 
is not necessary to describe, as if many- watches shonl 
ever be made on this principle it would be betU 
to adopt one form and keep to it. Mr. Eullbet^s i 
said to answer very well, and in one or two instanc* 
where I have applied it I have found it to do so. Mi 
Chalfont's is very ingenious and seems strong, but 1 hat 
had no practical experience of its action. In nearly a 
the fusee keyless watches I have made I have adopte 
the compound fusee wheel, and have never found i 
to fail when properly made ; ihe action does not reqnir 
to be shut off by the case, and the 'winding is smoot 
and easy ; but it cannot be employed in a small fla 
watch— at least, it would be unwise to do so^ 



185. The First Pendulum Clook.— It is somewhi 

uncertain who was the first to apply a pendulum to 
clock, the merit of its application having been claime 
by Dr. Hooke, Huygens, Galileo's eon Vincenzo, as 
others, but it is generally agreed by writers on tl 
subject that the first clock with a pendulum m 
made by a London clockmaker named Richard Hair 
for the old St. Paul's Church, Covent Garden, in 164! 

Chap. XX.] PENDULUMa 275 

or eight years before Vincenzo Galileo claimed to have 
adapted the pendulum to a clock. This is attested by- 
Thomas Grignion, of Great Russell Street, Covent Garden, 
the maker of the present clock, who had a brass plate 
engraved with an insciiption to that effect placed in the 
vestry. This inscription states that the said clock was 
the first long pendulum clock in Europe. That Grignion 
and his father had every opportunity of inspecting the 
clock before it was destroyed when the church was burnt, 
in 1795, there can be no doubt, but that he made a 
mistake, either in the date of its construction, or in 
stating that it had a long pendulum, is equally certain, 
for such a pendulum could not have been applied with 
the escapement then used. 

This has been pointed out by the Rev. H. L. Nel- 
thropp in his valuable chronological Treatise on Watch- 
work. After carefully summing up the arguments for 
and against the claims of Harris, Mr. Nelthropp says : — 

" 1st, That credit ought to be given to Huygens, 
the Dutch astronomer, for being the first to apply the 
discovery of Galileo to a clock; 2nd, That Dr. Hooke 
can justly lay claim to having brought the whole matter 
to perfection by the invention of his anchor escapement, 
which enabled him to use a long pendulum with a heavy 
bob, thereby rendering the arcs of vibration shorter, and 
necessitating much less motive power." 

186. Isochronism of the Pendulum. — Soon after 
the employment of the pendulum for the purpose of 
regulating clocks, Huygens investigated and estab- 
lished the theory of its isochronism. He discovered 
that in order to perform long and short vibrations in 
equal times, a pendulum must describe such a curve that 
the distance of the mass of a mathematically simple 
pendulum from the point of rest shall be always propor- 
tional to the force acting upon it ; and he proved that 
this condition is only secured by a pendulum oscillating, 
not in a circular arc, but in a cycloidal curve. In 
order to compel the bob to traverse that curve, he 




invented tlie cycloidtd cheeks — i.e., two metal cheek: 
placed one on esoh aide of the Hnspeiudoii spring of ib 
pendulum (which he made very thin, so that it woul 
bend eaaily to the form of the che«^), of the aht^ 
evolved from the cjcloid which the bob woold tract 
But, for reasons I have already given, these failed t 
secure the end for which they were designed, althoug 
they were a good deal fancied for a time, 

167. The CiTonlar Xiror.— A theoretically simpi 
pendulum consists of a heavy weight, having no magn 
tnde, suspended by aiid capable of oscillating at the eo' 
of a rod or cord having no weight 

The discrepancy between the time of a pendulni 
vibrating in e. circular arc and its true theoretical pat 
ia known as the " circular error." 

This error must have been of coosiderable importsiu: 
in the clocks constructed at tiiia period, with thei 
crown wheel escapements, light pendulums, and Ion 
arcs of vibration ; but subsequently, upon tha iatrodii( 
tion, about 1680, by Dr. Hooke, of the recoil eecapemenl 
with a longer and heavier pendulum, and consequen 
shorter arcs, the difference was found to be hardl, 
observable. In fact, it is actually useful in connteractin 
a tendency to gain in the long arcs which exists in thi 
form of escapement ; while in the dead-beat escapemen 
the arc is so short that the actual and theoretical padi 
are practically identical 

188. Time of One Swing of Simple Fendnlnm.- 
The time of vibration of a pendulum varies with ih 
square root of the length, and the following is tii 
mathematical formula usually employed for fimli-ng tii 
time of one swing of a simple pendulum. 


in which ir is the symbol used to express S'1416, whic 
is the ratio of the circumference of a circle to its diantetei 
/ is the lei^jtlk of the pea&\:^u'aL m ^£«t, and g the accleri 

Chap. XX.] PENDULUMS. 277 

tion due to gravity of a body falling in vacuo, which is 
3 2 '2 feet per second in the latitude of London. Substi- 
tuting these values in the above equation with one second 
for t, we find the length of the seconds pendulum to be 

^^'^ or 3-2616 feet 


The value of g varies slightly in different latitudes, 
increasing with the latitude and diminishing with the 
altitude of the place, and consequently the length of a 
pendulum to oscillate in any given time also varies. 

The length of a simple seconds pendulum, i.e., a 
pendulum taking one second of mean time for each 
swing, in the latitude of London being 39 '1393 inches ; 
it is 39 inches at the Equator, where the value of ^r is 
less, and 39*206 inches at the Poles, where it is greater. 

The length of a pendulum for any required time may 
be found, therefore, by multiplying 39*1393 inches by 
the square of the time required. 

Example 1. — ^To find the length of a pendulum to 
beat 3 seconds : 

32 = 9, and 39-1393 x 9, 
= 362*2537 inches, 
= 29*354475 feet, the length required. 

Example 2. — ^To find the length of a pendulum to 
beat f of a second : 

^3 = 1 and 39-1393 x | 

= 17*3952 inches, the length required. 

To ascertain the time in which a pendulum of any 
given length wUl swing, we have the proportion 

in which L is the given length, and 

Xf or Vx'^ is the time iq(\>\\£^. 

Example 1. — Find the time in. which a peiidulu 
meaauriiu; 22 mchea will swing : 

Example 2. — Find the time in which a pendula 
measuring 3522537 inches will swing : 

These calculations give the theoretical lengths onl 
of pendulums, i.e., the actual lengths of mathematical! 
simple pendulums (the nearest approach to which 
practically a boh of lead or other metal of great specif 
gravity suspended on a fine thread of some sort), but i 
practically the rod has to be made thick enough to ensni 
sufficient rigidity to resist beading and quivering vbe 
it receives impulse, and a heavy bob must be of conside 
able size, the efiective centre of the swinging weigh 
corresponding to the weight of a simple pendulum, i 
always situated at some point above the extreme end ( 
an ordinary pendulum. 

189. The Centre of OBcillation.— This point is calle 
the "centre of oscillation"; it is, in fact, the centr 
of percussion, that is to say, the point at which a foive o 
resistance must be applied to stop the pendulum suddenl 
without jarring the other parts or producing any pressut 
on the point of support, or, more properly speaking 
destroying its equilibrium. 

This point does not correspond to the centre ( 
gravity of the mass of the pendulum, which ia a fixe 
point, but ia always a little below it, but in a long au' 
heavy pendulum with a light rod it is situated so nea 
that point that its position may be approximately detei 
mined practically by placing tho rod on a knife edge an' 
shifting the bob till it balances at the required length 
this operation, will ^ve tbe centre of gravity, and the bo 
wiJl de found to requke vei-i ^iViXft ■re^'AaJCoiJt ui^Kt^ards 

Chap. XX.] PENDULUMS. 279 

From the want of uniformity in the shape of a com- 
pound pendulum it is diflScult to obtain the radius of 
oscillation mathematically ; it is done by dividing the 
sum of each particle of the mass multiplied by the 
square of its distance from the point of suspension by 
the sum of each particle multiplied into its distance. 

The distance between the centres of "suspension and 
oscillation, or the length of a simple pendulum, to beat 
any required time may be found experimentally by 
means of a pendulum having an axis at each end, one of 
which must be adjustable, thus utilising the fact that the 
centres of suspension and oscillation are convertible. 

When the pendulum has been adjusted to oscillate in 
the time required, by moving the weight to the other 
end, inverting the pendulum and adjusting the axis 
until it beats in the same time, the distance between the 
axes will be the theoretical length of the simple pen- 
dulum for that time. It was with such a pendulum that 
Captain Kater made some of his gravitation expeii- 
ments, for which purpose it is still used, but this is not 
what is called Kater's pendulum, which is a mercurial 
pendulum with glass jar and rod. The length of a side- 
real seconds pendulum in the latitude of London is 38*87 
inches ; it varies in the same ratio as the mean time 
pendulum, and its time for various lengths, and length 
for various times, may be calculated in the same manner 
by substituting 38'87 for 39-1393 inches as a datum. 

1 90. Suspension of Pendulum. — Soon after the intro- 
duction of heavier pendulums it was found better to 
suspend them separately and to connect them with the 
pallets by passing the rod through a bent piece called the 
crutch, which communicated the impulse to the rod. This 
had, indeed, been done by Huygens, who hung his pen- 
dulum on two cords between the cycloidal cheeks, but 
these cords were found to possess disadvantages in the 
shape of insufficient rigidity (and of friction, even when 
the cycloidal cheeks were abandoned), and were discarded 
in favour of the suspension spring. GVookxcaksets^^^^scis. 


280 WATOB AMD CLOCK MiKtSa. [Cbrnt-U 

SO far, when tha dead beat eoc^Mment was introdaced, a 
to maintain that they could Becure isochronism in the r 
brations by adopting certain lengths and thicknesses ( 
this spring, tapering it, &c. However, these theories hsT 
Ion;; since been exploded, and clockmakers now trust t 
obtain isochronism by constructing their escapements o 
scientific principles rather than by interfering with & 
natural harmonic motion of the pendulum. These mil 
conceptions arose chiefly from the castom, still oommo 
amongst a good many watchmakers, of confounding ti 
action of the pendulum in its relation, to gravity wit 
that of a balance of a watch and its spring, betwee 
which there is no analogy, the conditions being entirel 

The suspension of a pendulum ia a very importat 
(actor in the performance of a clock, the desideratoo 
which has now for a long time been recognised, being t 
obtain the maumum resistance to bending and quiyerio 
when the impulse is communicated to the rod, with tli 
minimum of friction and resistance to the free action c 
the pendulum. 

(With a correctly constructed dead-beat escapemei 
the vibrations of the pendulum are practically ix 
ohrououB up to 6° of arc.) 

Since this principle has been fully recognised, varioo 
methods have been devised for suspending the penduloi 
so that its free vibration shall not be interfered witl 
Amongst others more or less ingenious, I may mentio 
the following contrivance by Vulliamy : — A gun-mett 
bar was fixed firmly to the top of the pendulum rod i 
'right angles to it ; this bar rested on a knife edge, ( 
knife edges, the greater part of the weight of the pei 
dulum being kept off the knife edge by four count« 
poising weights which were attached to the bar by core 
that passed over pulleys fixed above the suspenAioi 
Many years ago a two-seconds pendulum with a vei 
heavy bob was suspended in th« way to a latj^ pnbl 
clock in Gla^ow \jy 'Beiqttnan. I^ufe, ^-^^^ Ss^^sDJai 

Chap. XX.] PENDULUMS. 281 

clockmaker, who had been for many years in Vulliamy's 
employ ; but with such an arrangement, the axis of sus- 
pension must be above the pallet axis, and the pallet 
and crutch will consequently move through a larger arc 
than the pendulum, which entails friction at the fork ; 
in addition to which the friction of the cords, &a, fully 
compensates for any advantages of a knife-edge sus- 

I believe the pendulum referred to has since been 
replaced by one with the ordinary spring suspension. 

Other contrivances were tried with more or less 
success, with the result that the spring has proved 
superior to them all, and it is now the only method of 
suspending the pendulum in use, with the exception of 
that employed in some of the small French timepieces, 
in which the crutch is prolonged to form the rod, and a 
knot screwed on to the end of this forms the pen- 
dulum bob. 

John Whitehurst, of Derby, suspended his pendulums 
from the pallet axis in large turret clocks, which answered 
very well with the escapement he used, which was the 
pin wheel. Referring to this, the late Charles Frod- 
sham, who from his many exhaustive experiments 
with the dead-beat escapement is a high authority on 
this subject, said he made experiments with the Graham 
escapement so applied, but with unsatisfactory results, 
and that doubtless the pin wheel was better adapted for 
this mode of application. 

The strength of the suspension spring, length of the 
crutch, and the length and weight of the pendulum, are 
directly dependent on one another. The axis of suspen- 
sion should be fixed so that the bend of the spring will 
be opposite the pallet axis. 

The crutch should be as short as possible in propor- 
tion to the length of the pendulum, so that it does not 
affect the spring, or cause it to bend in the reverse 
direction to that in which the pendulum is travelling, 
when the impulse is communicated tA ^Vi^ xcA ^^<^i*s^^^ 


the fork, as waa sometimes noticeable in the very loi 

SndulumB formerly applied to the old turret clock 
lis qnivering of the spring at the moment i^ tl 
impulse being given is seldom observed now that it hi 
become a rule never to apply a pendulum of more tb 
two seconds. 

The 20- or 30'feet penduluma were formerly appBe 
to turret clocks with the idea of better oveTcomin 
irregularities of impulse, ius. They are, however, nen 
used now ; the shorter pendulums require less impnli 
and less compensation, and are less liable to disturbuu 
from the external inQuences of wind and atmoepheri 
changes, not to speak of the difficulty of finding room fi: 
the long pendulum in the turret. The necessaiy contn 
over the clock is now obtained by using vety he»v 
pendulum bobs. 

For regulators, the on&seconda pendulum is nevt 
exceeded, and is generally adhered to, being a ooi 
venient length ; but of course any decrease of the lengt 
of a pendulum is attended with a greatly increased e^ 
of errors in the escapement, shake in the boles, &c. 

191. Pendulum Cock. ^Another point of tiie ma 
vital importance to the good performance of a clock, an 
one that is not always carefully attended to, is tl 
manner of fixing the pendulum cock or bracket. I 
turret clocks with a heavy pendulum bob this shonl 
either be built into the wail opposite the clock, or bdb 
through it in a very firm manner, as if this is not carefnll 
attended to the centrifugal force of the bob will cause 
slight racking motion of the bracket to take place, ani 
however small the arc, a slight falling off in the extent i 
the vibrations will be observable. If the spring be m 
securely fastened at the end to the cock or bracket, tl 
same efiect will of coui-se take place. 

The pendulum cock of a regulator should be secure 
bolted to the back of the case, which should not rest ( 
the ground, but be fixed to a permanent "wall of ti 
rooia, and the baak sfao'iiS. Nw bku-^sMi-j ■e«isHbw(atJai__i] 

Chap. XXI.] PENDULUMS. 283 

less than 1 J inches in thickness. Keid recommends that 
a bracket be in this manner fixed to the wall, and the 
movement screwed or bolted to it, the front of the case 
being capable of being pushed on or pulled oflf ; and this 
is the method adopted in all the best regulators, which 
are in addition fitted with an inner case, baized at all the 
edges and crevices, to exclude dust. 

There is another error of the pendulum caused by 
the variation in density of the atmosphere, and con- 
sequently in the resistance to the pendulum's motion 
through it. 

192. Pendulum Bobs. — ^Yarious-shaped bobs have 
been from time to time devised, the most common in all 
the old clocks being the lenticular or lens-shaped which 
has been recommended by Reid and others. If exactly 
made this would not be a bad form for the bob, but it is 
a very difficult one to form accurately, one of the convex 
sides being invariably more protuberant than the other ; 
or, if these are properly divided, there is another difficulty 
in getting the hole exactly in the middle of them. Any 
inequality of this kind will cause the pendulum to have 
a twist at every swing, and prevent the good going of the 
clock ; for this reason pendulums are now nearly always 
made with cylindrical-shaped bobs, which is also a con- 
venient shape for compensation pendulums. 

193. Barometric Error. — The error caused by varia- 
tions of the density of the atmosphere is called the 
" barometric error," and various methods have been tried 
for its correction, such as causing the pendulum to vibrate 
in a vacuum, fixing small barometers to the pendulum 
rod, (fee. ; but the first of these could hardly be adopted in 
turret clocks, and in the latter the adjustment is very 
troublesome. The best plan is to make the pendulum 
describe so large an arc that the circular error will cor- 
rect the barometria This has been adopted by Sir E. 
Beckett in the great Westminster clock, in which the 
pendulum describes an arc of 5^ 30' at each swing, 
which is found to absolutely correct it. 


104. Bognlfttion. — The usual mode of regolating pe 
duluDu is by a nut at the end of the rod, by the screwi: 
up or letting down of which the pendulum is mode 
vibrato faster or slower as required, and this is a sui 
cieutlj efficient air&ngement for ordinary house doc) 
But iu clocks where it is difficult to get at the penduln 
to perform this operation, such as quarter clocks, the r^ 
lation is usually effected by turning an arbor, the sqna 
of which proJectA through a hole in the dial, at the oth 
end of which a pinion ia fitted, which gears into anoth 
pinion or wheel fitted to a screw, which is fixed to tl 
top of the suspension spring in such a manner that ! 
turning it it raises or lowers the spring through the chof 
altering the effective length. This is the method of reg 
lating moat of the ornamental French drawing-roa 
clocks. But for turret and astronomical clocks the fit 
method is objectionable from the necessity of stoppii 
the pendulum in order to regulate it^ which ia found ' 
affect the rate, while in the Litter it is difficult to fit tl 
spring in the chops so nicely that it can be drawn np 1 
tike screw and down by the weight of the pendulum, ai 
be at the same time so rigid as not to allo^r of any pli 
taking place. The general way of effecting the temponu 
regulation of turret clocks is by placing small weighta( 
a platform left for the purpose some distance up Uie n 
above the centre of oscillation, adding to them to bm 
lerate and taking from them to retard the vibr&tioi 
la a compound pendulum with a conioal-shaped boh tl 
top of the bob should be tapered up to a oone. It k 
flat the dust which will accumulate will accelerate t 

Chap. XXI.] 285 



195. Ratio of Expansion of Metals. — The following 
is the expansive ratio of the different metals capable of 
being used in the compensation of pendulums, with regard 
to one another, taking mercury as the unit : — 

Mercury (in bulk) '000 
Lead . . • '176 
Zinc . . . -170 

Tin . . . -117 
Brass . • . *100 

Iron (wrought) • "070 

Iron (cast) . . '066 

Steel (soft) . . -062 

Steel (tempered) . '065 

Glass . . '046 

196. Mercurial Compensation Pendulum. — Graham 
was the first to conceive the idea of counteracting the 
effects of the lengthening and shortening of the pendu- 
lum by the expansion and contraction of the material 
composing it in heat and cold, by utilising the known 
different expansibility of various metals in its construc- 
tion, by means of which the centre of oscillation could be 
kept at the same distance from the point of suspension 
and the time of the vibrations constant. This idea 
resulted ultimately in his invention of the mercurial 
pendulum, of which Fig. 62 is one of the most approved 
and generally constructed patterns for regulators and 
astronomical clocks. M is a glass jar containing mer- 
cury, s the sole of the stirrup, countersunk, on which 
it rests, T is a metal top fitted tightly over the jar to 
exclude dust, etc., the projecting portions of ^which 
embrace the sides 8 8 of the stirrup, and K is the nut at 
the end of the rod r, for regulating the final adjustment 
of the length of the pendulum. The point attached to 
the sole s indicates on the degree-plate fixed to the back 
of the case the arc through which the pendulum oscillates, 
while the point c attached to the top of the stirrup frame 
points to Uie divisions marked on the timing nut K, by 


Hliicli llic rcgiilntinn in fiicilitatt^. 

Frodsliani recommended a 
jar iiisteatl of a glass on 
Ifing a better coiidui'tor, 
tlie rod ilijipiiig into the 
curv, "so that the nbc 
aluiust siiiiultaneouslv aff 
by cliaiigi-s of ti'inperat 
111 i-eference to tLe ste* 
Iti-id snys : " A. jar of 
kiud, fi-om its Ijeing 
tliin (for it would be 1 
were it as tliick as a 
one), would be easily afl 
liy tlio clianges of tempen 
and mercury being still 
susceptible of these chi 
the operation of eountem 
the etTeeta of them mig! 
too Biidden," in additic 
ixiinting out the danger 
steel jar becoming mognel 
but this, ev<>ii did it t 
would not affect the goinj 
clock with a properl 
pended pendulum and, 
that it is admitted to be 1 
to use considerably lieaviei 
than were customary in ] 
time, the relative weightsi 
jar and its contents wou 
much tlie sanie. There is 
doubt that for the 401b. 
dulum spoken of by Sii 
muad Beckett, glass jars " 
^'*- '^r^^lwS'i.S""™"" be utterly unsuitable, as 
would be great ditficuV 
getting them made perfectly true and fixing thf 


the pendulum so that the mass would be proportional on 
both sides of the plane of its vibration, to say nothing of 
the diflficulty of handling so great a weight in a glass jar ; 
and for such pendulums, where mercury is employed as 
the compensator, steel or, still better, cast-iron jars, should 
be used (brass, or other metal than iron or steel, could 
not be employed, as the mercury would in time dissolve 
it) ; but for regulators and other good clocks where this 
form of compensation is adopted, with one-seconds 
pendulums, there is no reason for changing what has 
answered so well and looks so hand soma 

In calculating the height of a column of mercury the 
expansion sideways of the jar has to be allowed for, but 
as any calculations for this purpose can at the most be 
only approximate, the final adjustments of the pendulum, 
both for compensation and length, requiring to be made 
after the clock is going, I will put down the proportions 
of a good mercurial pendulum. 

Total length from the top of acting part of the 

spring 44-0 

Height of jar 8*0 

Inside diameter of jar 2-6 

Height of column of mercury . . . 7 '6 
Weight of ditto . . about 14lh. 

The rod and stirrup frame should be of suitable sub- 
stance and rigidity, and may be of either round or square 
steely as, although some stress has been laid on the 
ad^'antages of the former over the latter shape for the 
rod, there is practically no difierence, and the shape may 
therefore be left to the taste or fancy of the maker. 

The expansive ratio of mercury to steel is about 
sixteen to one, and, with the dimensions given, the rise 
and fall of the mercury in the jar, allowing for the side 
expansion of the jar, are nearly six times as much as the 
lengthening and shortening of the pendulum in heat and 
cold, which, subject to slight final adjustments, will 
maintain the radius of oscillation constant. 

The screw upon which the regulating nut, K, is fixed 


has thirty threads to the inch, and the nut is divided 

into thirty, each division being equal to about a second 

in the daily rate of the clock. Formerly, clockmakers 

I had always to add to the height of the mercury deduced 

from their calculations in order to obtain sufficient com- 
pensation, which was attributed to the extra stiffiiess of 
the pendulum spring in cold, and vice versd. This, as 
Sir Edmimd Beckett observes, was a most inadequate 
cause, considering the thickness of the spring in relation 
to the weight of the pendulum, and he points out that 
it resulted from a mistake originally made by Francis 
Baily, P.R. A.S., in neglecting the weight of the jar itself 
in his calculations. Sir Edmimd says : — ** The weight of 
the glass jar and rod may be a sixth of the ultimate or 
corrected weight of the mercury, and about so mucli 
more height must be added for it, as it is evidently much 
the same as if all the* bob of pendulum were mercuiy, 
but its rise only five-sixths of its real amount. And 
approximately that is near enough, seeing that pendu- 
lums vary in their proportions, and must be finally 
adjusted by trial" 

197. The Gridiron Pendulum was invented about the 
year 1726 by Harrison. It consists of nine rods, five of 
steel and four of brass, so attached to cross pieces that 
they expand and contract in contrary directions, and so 
keep the bob at the same distance from the suspension. 

The co-efficients of expansion of the metals usable for 
pendulums, or the fraction of a foot which a rod, which 
is one foot at the fireezing-point^ expands for a rise of 
1® F. are as follows : — * 

Glass . . 


Brass • . -0001042 

Platinum . 


Silver . . -0001068 

Steel . . 


Lead • . -0001640 

Copper . . 


Zinc . . -0001663 

* From recent experiments it has been found thai the oo-effideiit 
of expansion for zinc generally given is too high. If the ratio in the 
above table, for example, were taken as the bans of calculation for the 
dimensions of a zinc and steel pendulum, we should have a length (A 
ano of 28*5 inches, which would be too great. 


The following are the lengths of rods prescribed by 
Cumming : — 

Outside steel rods from pin to pin . . . 29*5 
Middle steel rod from the upper end of pen- 
dulum spring to pin at lower end . .31*5 
Inside rods from pin to pin . . . . 24*5 
From the pinning of the lower end of outside 

steel rods to the centre of the ball . . 5'0 

Total length of steel rods . • . 90*5 

Outside brass rods from pin to pin . • . 26 87 
Inside do. do. ... 22*25 

Total length of brass rods • . 4912 

Reid found, however, that with these proportions the 
pendulum was rather undercompensated, and recom- 
mended a total length of steel of 112-568 to 71 inches of 
brass. The total lengths should be inversely proportional 
to the coefficients of expansion for the metals used. 
These pendulums have now for a long time been quite 
superseded by the mercurial in regulators, and are only 
to be met with in old clocks. 

A cheap and effective form of compensation is the 

198. Wood and Zinc, or Wood and Lead, Pendulum. 
— A wooden rod having a bob of lead, or zinc, resting 
on a nut at its end, the expansion of the bob upwards in 
heat compensates for the elongation of the pendulum, and 
vice versd. Zinc is the better metal of the two for the 
bob, its expansion being nearly as much, while it is not 
BO likely to be dented and bruised in handling as lead. 
Objections have been made to the wooden rod on account 
of its liability to shrink and warp with atmospheric 
changes, but this may be to a great extent prevented by 
using well-seasoned wood thoroughly dried and varnish- 
ing it. 

Zinc expands in heat about three and a half times as 
much as deal, and the following are very good propor- 


tions for a one-seconde zmc oompeosation pendulnm «i 
a deal rod ; — 

Len^ of pendulum from Uie top of snspen- 

siOQ Bpriag to the bottom of the bob . . ii-6 

Or CO the bottom of the nut . . . . 45'2S 

Height of hob 10-6 

Diameter of ditto 2'0 

Weight of ditto, 8 lb. 

Acting length of lUEpenaion Bpring . . 1-0 

Width of ditto -46 

ThickneBS of ditt<i -008 

Diametet of rod -6 

The top of the rod has a brass collar fixed on it, reachii 
five or six inches down the rod, the upper part of whi' 
is solid with a slit in it to receive the pendulum spring 
199. ZiRC and Steel CompenBation. — ^The first to a 
ploy zinc as a compensator was John Smeaton, F.R! 
the engineer who built the Eddjstone Lighthouse. Asl 
pendulum had a glass rod and a lead bob which expandi 
upwards, the length of zinc required was very ntuch le 
than in the zinc and steel pendulums with cast iron bol 
suspended or attached to the rod at their middle, as > 
present used for turret clocks, for which the glass rc 
would, of course, be inapplicable. 

For all very large clocks this method of compens 
tion is perhaps more suitable than any other, as it 
comparatively cheaper and is sufficiently aocnrate. 1 
fact, Sir O. £. Airy Etat«d that in the standsi 
sidereal clock at Greenwich, to which it was applis 
it answered quite as well as a mercurial pendulun 
which should never be employed where the bob 
reiiuired to exceed twenty pounds in weight. 

The construction of the zinc and steel pendulum is i 
simple that I have not thought it neoessaiy to give 
figure of it It is comprised of a hard-drawn zinc tul 
fitted accurately (but not so as to bind it) over the pe 
dulum rod, resting on a nut at the end. Enclosing tl 
zinc, and fastened to it at the top, is another tube < 


iron or steel, at the lower end of which the bob is fixed. 
To prevent the bob and tubes from turning when the 
pendulum is adjusted, a flat part of the rod is filed away 
and a pin is fixed through the zinc tube. 

The expansion downwards of the iron tube being the 
same as th6 expansion upwards of the bob if it is iron, 
the lower end of the bob may rest on the nut at the 
bottom of the tube without altering its relative expan- 
sion to the zinc compensation ; but if a lead bob be fixed 
in this manner the difference of its expansibility to that 
of iron or steel must be taken into consideration from 
the bottom of the bob to the centre of oscillation. The 
best way, however, is to attach it at the centre of oscilla- 
tion, by turning out a larger hole up to that point, so 
that the bob wUl slip over the nut at the end of the iron 
rod and rest on the seat so left. This plan affords a firm 
bearing, renders the amount of compensation more easily 
calculated, and obviates any discrepancies which might 
arise in the compensation in sudden changes of tempera- 
ture from the slow action of the large body of metal as 
compared with that of the rod, &c. 

To calculate the lengths of zinc and steel we have the 
proportion : 

The expansive ratio of zinc : the expansive ratio of steel : : 
the length of the latter : the length of the former. 

For a one-seconds pendulum, taking its length of rod 
and spring approximately at 45 inches, and the coefficient 
of expansion as in the ratio of 170 to 65, we have : 

170 : 66 : : 46 : «, 

.*. — TTT — = 17*2 inches ; 

this we may take roughly as the length of the iron tube, 
which we must now compensate for 

170 : 66 : : 17*2 : a?, 
66 X 17-2 

■-■-m- = '-'' 

and 17*2 + 6*6 = 23*7 inches. 


With this length it is foimd the pendulum is sligbt 
iindei'campensated, and it ia usual to allow 25 inches f 
the length of the zinc tube. The result is not eaej 
get at correctly, on account of the position of the cent 
of oscillation varying in pendulums of different-shap 
bobs, &c., so that it can only be arrived at approxitnatel 
even mathematically. 

Supposing, however, the bob to be attached at ti 
middle, which we know to be practically the centre 
oscillation, and assuming the zinc and iron tabes to be 
the same length : 

Let ( ^ the len^h of the steel rod from the point <d • 
pension to the centre of oacillation, 
„e ^ the length of the lanc tube, 
„ k ^ Uie length of the iron tube. 

Then tating the numbers, as in the previous propc 
tion, from the table of expansions, we have t^e equatioi 

■1701' = -065 I + 070*, 
whence -085 /— (170 — -070) o, 
and aBBunung / to he 39 incbeo, 
■065 X 39= U, 

and B ^ 26-36 inches, 
, . -^ = the length at sine required. 



200. Recoil Escapement.— Shortly aftor the applie 
tion of the pendulum to clocks, tbe want <rf a betb 
escapement became universally felt, owing to the irtwnli 
performance of those with the escapement then uw 
and the clockraakers of the day were directing » 
their attention to that point. It was for the purpose i 
correcting the want of isochronism in the vibrations < 


the pendulum that the cycloidal cheeks (already de- 
scribed) were devised by Huygens, but with no good 
results, as I have before stated. 

About the year 1666, Dr. Hooke invented what is 
known as the recoil, or anchor, escapement. This was 
afterwards adopted in a clock made by Clement^ a 
London clockmaker of the time, and, from the good 
performance of the clock, it speedily became generally 
made, entirely superseding the crown wheel escapement 
hitherto used. 

This escapement was the first step in the direction 
of securing isochronism in the vibrations of the pendulum, 
as it involved a longer pendulum, shorter arcs, a heavier 
pendulum bob, and less motive power. Consequently, 
this combination resulted in the pendulum being less 
controlled by the escapement, and therefore less influ- 
enced by variations in the impulse, although the escape- 
ment cannot be considered detached in the sense that 
a dead-beat one is. 

The recoil is the escapement used for most English 
clocks with short pendulums, for which it is well adapted 
where no very great accuracy is required ; it is easily 
made, and performs regularly where a fusee is used. 

But although variations in the impulse produce less 
alteration in the arc of vibration than similar variations 
would in the arc of the Graham escapement, which for 
some time led clockmakers into the belief that it was the 
more reliable escapement of the two, they affect the 
time of the vibrations very considerably (the clock 
going faster for any increase of the motive force, and 
slower for a decrease), as should be patent to any one 
without further demonstration, after a little considera- 
tion of the form of the pallets and the direction of the 
forces. Yet after the many years during which the two 
escapements have been tried, and the experience which 
has proved undeniably the superiority of the dead-beat, 
people may still be heard to assert that the recoil is 
the better escapement of the two. 


In Clement's escapement the entrance pallet wi 
convex and the exit one concave, and they were a&e 

wards made flat, but in both cases were found to ci 
away very fast^ owing to the friction when the recc 
takes place ;, to prevent Una, they were subsequently mat 

Chap.XXn.] ESCAPEMENTS. 295 

both convex, as shown in Fig. 63, which lessens the angle, 
and consequently, the friction at the recoil. 

The action of the escapement is shown in the figure, 
the tooth at f having escaped the pallet, the tooth on 
the opposite side has fallen on the pallet at E, when, 
having sustained a recoil, it will give impulse through 
the angle shown by the lines e o, a o, of 4° on the arc 
described by the pallets ; after it has escaped the comer 
a, the next tooth on the right-hand side will fall on the 
pallet, f^ which will have intersected the circumference 
of the wheel 4*^ in like manner. 

These escapements are sometimes very badly planned, 
the pallet staff often being planted without any refer- 
ence to the size of the wheel and pallets, and the angles 
of the impulse curves being afterwards formed according 
to the fancy of the workman, with equal disregard to 
the position of the two centres or the number of teeth 

The intersection of two tangents to the wheel at the 
arc of intersection of the pallets at its circumference will 
give the exact point for the pallet holes, and lines drawn 
across the curves, as at e a d,h f^ should be tangential 
to a circle, h d, having a diameter equal to the dis- 
tance apart of the wheel and pallet centres, when the 
pendulum is hanging at rest. 

201. Pin Wheel Escapement. — This escapement (Fig. 
64) was invented by Lepaute about the year 1 760. It is a 
very good and simple escapement for large clocks. The 
impulse is communicated to the pendulum through the 
pallets M and N by the semicircular pins which project 
from the plane of the wheel. The i*esting faces of the 
palletiS are arcs of circles struck from the centre o. 

The impulse angle is dependent on the length of the 
pallets from their centre and the angle of their impulse 
faces; the amount of impulse found practically to be 
sufficient is about 4^. 

The thickness of the pallets and one pin together 
should equal the space between two pins less the neoes- 


nary freedom, which ia very little ; and the distance &pBi 
of the estreme pointa of tha pallets ahould equal half th 
diameter of the pins. 

A line drawn from the centre of the vei^ npo 

Fis. 6*.— Pin Wheal BsoLpemsBt, 

which the palleta are fixed through the centre of tl 
pin which is resting on the locking face of the longe 
pallet, ahould be tangential to the wheel at that poin 
Lepaute'e eecapement, in order to ensnre thia tangentii 
action and to render the impulse equal, had pallets < 
equal length acting on opposite sides of the wheel ; bi 

Chap.XXn.) ESCAPEMENTS. 297 

this arrangement requires double the number of pins,, 
and is not necessary so long as the longer pallet is placed 
to act on the inside of the pins. Some of his escapements 
were made with the alternate pins in two rows, each 
row being a little nearer to the pallets upon which they 
acted than the centre of the action. 

Sir E. Beckett improved this escapement by cutting 
away a part of the front of the impulse pins, as shown 
at L, Fig. 64 ; this permits of shorter pallets, and renders 
the action safer. 

The pin wheel, although a very good escapement for 
turret clocks, has been quite superseded by the gravity, 
and is inferior to the Graham, for regulators and 
smaller clocks. Very few of them are now made. 

202. Dead-beat (G-raham) Escapement. — This escape- 
ment was invented by Graham, at the end of the 
seventeenth century, and is the one generally applied 
to all the better clocks with seconds pendulums, for which 
it is best adapted. It was the direct outcome of Dr. 
Hooke*s recoil escapement, after the capabilities of that 
escapement were known, to which it is infinitely superior 
for the clocks to which it is applied. Fig. 65 shows the 
proportions which are at present adopted in all the best 
escapements. The pallets are made to embrace eight teeth 
of the escape wheel, and, as in the lever escapement, 
may be made either with equidistant lockings, semi- 
equidistant, or circular, as in the figure. The locking 
faces are made concentric with their axis. 

The pallets of this escapement were foimerly made to 
embrace as many as fifteen teeth, but as it was found 
that after the clock had been going for some time the 
thickening of the oil on the resting faces affected the rate, 
it was deemed advisable to reduce the run .on these as 
much as possible, and pallets embracing fewer teeth were 
adopted, with the result that the present proportions were 
foimd practically the best, as when made closer they 
were not safe, owing to defective workmanship, shake in 
the holes, &c In the best clocks the pallets are always 

298 WATOH AHD CLOCK lURixa. i;c^.zx] 

jewelled, m &re the pivot holes of the atoff Vh 

Tig. 8E,— I><*d Beat {Ordum) EKapement. 

pelletB ranbradng eaght teetb of ■ thirty-toothed -whe 

Chap. XXn] ESCAPEMENTS. / 299 

the proportions of the pallets and position of the staff 
holes with regard to the wheel are readily determined. 

The wheel teeth are made as thin as is consistent 
with due strength, and are cut away at their faces about 
12® from the radial, so that their points only touch the 
resting faces of the pallets, and the pallets are made half 
the thickness of a space between one tooth and the next, 
less the drop, which should only be just sufficient to 
insure freedom. 

The escaping arc depends upon the length of the 
pendulum ; with a one-seconds pendulum 4® are usually 
allowed, 2® on each pallet, which is divided into IJ® of 
impulse, and ^® of rest. With an escapement of the 
above proportions, radii intersecting the centre of the 
impulse planes will form an angle of 90®, and tangents 
to these radii at the circumference of the wheel will 
form a square ; the point where they intersect being the 
position of the pallet staff holes, which may also be deter- 
mined by taking the distance of the chord of the arc of 90®, 
at the circumference from the centre of the escape wheel. 

Mr. Frodsham said that when the dead-beat escape- 
ment has been mathematically constructed, and is strictly 
correct in all its bearings, its vibrations are found to be 
isochronous for arcs of different extent, from 0*75 of a 
degree to 2*50 degrees; that injurious friction does not 
then exist ; that the run up on the lockings has no 
influence, nor is there any friction at the crutch ; that 
oil is not absolutely necessary except at the pivots ; and 
that there is no unlocking resistance, nor any inclination 
to repel or attract the wheel at its lockings. 

Professor Airy, in a paper on this escapement, says 
there is no friction when the arc of escapement is equal 
on both sides ; and this has been construed by some of 
their advocates as an argument in favour of equidistant 
lockings, and by others of equidistant impulse planes or 
circular pallets, but with no reason, as it applies solely to 
the angular velocity of the pendulum at each vibration, 
with regard to which Mr. Frodsham says ; — " Much 





difference of opinion has been expressed upon the con- 
struction of the pallets, as to whether the lockings or 
circnlar rests should be at equal distances from the 
pallet axis, with arms aad impulse planes of unequal 
length ; or at unequal distaiicea from the pallet axis, witli 
arms and impulse planes of equal length. In the latter 
case the locking on one side is 3" abore^ and on tlie 
other 3" below the rectangle, whereas, in the former, tbe 
tooth on both sides reposes at right angles to the line of 
pressure, but the length of the impulse planes is imequaL 
When an escapement is correcUj made upon either plsn, 
the results are very similar, though I decidedly prefer the 
pure right-angled locking, although the arm of one pallet 
is longer than the other by the thickness of the pallet 
The angle at the tooth will, however, be the same." 

As in the lerer escapement, the balance of opinion is 
in favour of circular pallets, and as there are no real 
advantages in the one form over the other in a properiy- 
constructed escapement (see pages 190-4), the differ- 
ence, if there is any, being so slight that it cannot be 
perceived in the results, the trade do right in adhering 
to the form which can be moat conveniently made. 

203. Conditions of Isoohronons Vibration. — Mr. 
Frodsham, referring to Mr. W. J. Frodsham's pen- 
dulum experiments, and Mr. TuUiamy's pamphlet on 
the dead-beat escapement, regretted that the first were 
made independent of the escapement, and that the latter 
did not connect it with pendulum experiments ; for, he 
says, " they cannot be separated, no matter the character 
of the escapement employed, whether gravity, remontoii, 
detached, or dead-bekt" He laid down the following 
conditions on which alone the isochronous vibrations of 
the pendulum depend ; — - 

" 1. That the pendulum be at time with and without 
the clock, in which state it is isochronoua, 'suspended by 
a spring.' 

" 2. That the crutch and palleta shall each travel at 
the same precise angular velocity as the pendulum. 

Chap. XXn.) ESCAPEMENTS. 301 

which can happen only when the arc each is to describe 
is in direct proportion to its distance from the centre of 
motion — ^that is, from the pallet axis. 

"3. That the angular force communicated by the 
crutch to the pendulum shall be equal on both sides of 
the quiescent point ; or in other words, that the lead of 
each pallet shall be of the same precise amount. 

" 4. That one or other number of degrees marked by 
the crutch or pallets shall correspond with the same 
degree or degrees shown by the lead of the pendulum, as 
marked by the index on the degree plate. 

" 6. That the various vibrations of the pendulum be 
driven by a motive weight in strict accordance with the 
theoretical law ; that is, if a 5 lb. weight cause the 
pendulum to double its arc of escapement of 1®, and 
consequently drive it 2°, all the intermediate arcs of 
vibration shall in practice accord with the theory of 
increasing or diminishing their arcs in the ratio of the 
square roots of the motive weight. 

"To accomplish the foregoing conditions there is but 
one point or line of distance between the axis of the 
escape wheel and that of the pallet, and that depends 
upon the number of teeth embraced, and only one point 
in which the pallet axis can be placed from which the 
several lines of the escapement can be correctly traced 
and properly constructed with equal angles, and equal 
rectangular lockings on both sides, so that each part 
travels with the same degree of angular velocity, which 
are the three essential points of the escapement." And 
again, " It is possible to obtain equal angles by a false 
centre of motion or pallet axis, but then the arcs of 
repose will not be equal ; this, however, is not of so 
much consequence as that of having destroyed the con- 
ditions Nos. 2, 3, 4, for even at correct centres, if the 
angles are not drawn off correctly by the protractor and 
precisely equal to each other, the isochronous vibrations 
of the pendulum will be destroyed, and unequal arcs will 
no longer be performed in equal times ; and I consider 



tliat the quieecent point ia not the centra of the Tibnlia 
except when the driving foroes are eqaal on both sides i 
the natural quiescent point of the pendulum at rest 

" Now this ia the very pith of the sabject, and vhit 
few would be inclined to look for with any hope < 
Ending in it the solution of this important question, tl 
isochronistn of the pendulum. 

" One would n&tutally suppose that unequal arcs ( 
the two sides of the vertical lines would not serioiul 
aSect the rate of the clock, but would be equal si 
contrary, and consequently, a balance of errors, and i 
they probably are for the same fixed vibration, but m 
for any other, because difierent angles are drawn wh 
different veiocities ; the short angle haa a quicker nte i 
motion than the long. Five pounda motive weight w\ 
multiply three times the pendulum's vibration over i 
arc of escapement of 0'75° ; but the aome pendulm 
with an arc of escapement of 1 **, would require 1 1 -20 1 
to treble its vibration ; whence I proved that the tiin 
of the vibration varied in the same ratio as the sum 
the squares of the diflerence of the angles of each paU( 
compared with the spaces passed over ; nor is this a va 
diflicult question to solve mathematically, nor difficult 
practical proof" 

204. Frodeham's Tables- — Mr. Frodsham gives tl 
following tables of the principal parts of the Grabs 
dead-beat escapement for aatronomical clocks, suited i 
an escape wheel of one inch radius and thirty teet 
for each of the several numbers of teeth that may 
embraced, from two teeth, or one apace, to thirto 
teeth, or twelve apaces ; " from which tables those for a 
other radius may be obtained by simple multiplication : 

"Table L — On the VuUiamy principle, with pal 
arms of equal length and circular rests, or lockings 
unequal distances. 

"Table II. — On the principle of circular rests 
lockings at equal distances from the pallet axis s 
unequal pallet arml. 


Graham Dead Beat Eace^mnetO. 














Dwt. • 












ti '■ 







13! ^ 







Chord S. 
inches. in 








1^0318 3 




ia always liirtaible by twi 

ccreeB in t} 



"It U impoBsible to examine l^esie tAbles without tli 
attention being specially called to the numben peculia 
to the plan of escaping over eight teeth apon t^e tn 
rest principle, in 'which the following facts appear : — 

" (1) That the angles at the wheel and p^et axw ir 
each 90O. 

"(2) That the radius of the circular locking is eqnslti 
the radius of the wheel 

" (3) That the chord of the opening of the wheel tni 
the distance of the pallet axis from that of the wheel *n 
each equal, and equal to the square root of the sum li 
the sqiinres of the radius of the wheel, and the radios d 
the pallet arm at its locking. 

" This number, therefore, embraces all the points in 
Euclid's forty-seventh problem — renders the escapement 
most easy to delineate and make a working skeUJi ; mcI 
were a diagram admitted here, it would probably win for 
itself the title of the geometrician'B number, and so fix it 
in the mind as to promote its universal adoption. 

"The calculations in the above tables are based opoo 
two slightly different plans, which may be thnt 
defined :— 

" In No. 1, the centre of the pallet axis, a point of ilu 
utmost importance to the correct gearing of the esc^ 
ment, is found in the point of meeting of the two chord) 
representing the thickness of the jmlleta prcJonged to 
their intersection ; or it may be otherwise described u 
the corrected tangent of the upper and lower rest. 

" In No. 2, the centre of the pallet axis is found ii 
the intersection of two tangents drawn from the pointi 
where the said radii of the wheel teeth meet the ciid> 
circumscribing them, a construction which makes tlu 
circular rests, or lockings, at equal distanoe from tbi 
pallet axis, and requires the pallet arms to be o 
unequal length. 

"In either case, when the centre of the pallet axi 
is found as here -directed, and all the several measnie 
given in the tables have been properly carried oul 

CSiap. XX 11.1 ESCAPEMENTS. 305 

the escapement will be found to possess properties and 
advantages which are not to be obtained bj any other 

''(1) The action of the tooth upon the inclined faces 
of the pallets will be uniformly the same for each 

" (2) If the line connecting the wheel and pallet axes 
be bisected, and upon the point so found a circle be 
described with a radius equal to half this distance, the 
circumference will pass through the centres of the two 
axes, and also through the locking points, as each tooth 
and pallet in action becomes alternately engaged ; thus 
proving that the locking points are angles in a semicircle 
and consequently right angles^ and aJso that the two 
semi-angles at the wheel and pallet axes will be equal 
to a right angle, and therefore complements to each 

"(3) The arc described by the pendulum and that 
described by the escapement will be perfectly equal, 
the result of which will be that the same amount of 
force being constantly transmitted through the escapement 
to the pendulum, it will be driven though various arcs 
in equal times, and the pallet, crutch, and pendulum, 
instead of interfering with each other's motion, wiU 
travel together step by step, with the precise angular 
Telocity due to their respective distances from their 
centre of motion. And because the angles at the lock- 
ings are right angles, and consequently perpendicular to 
the action of the main force, there will be neither -draw 
nor repulse." 

Theoretically, the smaller the escaping angle of this 
escapement) and consequently the run on tiie dead faces 
of the pallets, the better, but it is found practically 
unsafe to make this much less than the angle I have pre- 
scribed, namely, 2^ from zero, with pallets esci^ing over 
eight teeth ; and even with this angle the work must be 
very accurately done and well finished, with no more 
shake than is absolutely requisite. 


Although Sir George Airy oonduded that tl 
friction on the dead ftwee of the pallets did not affect tl 
rate of clocks with this eecapement, there is no don] 
that any chsnge in the oonaisten^ of tlie tnl on the 
faoea wUl affect it, and Mr. Bloxam, the inventor (d tl 
dipleidOBCope, attributed all the errors of the escspema 
to this friction. Sir E. Beckett says :—" Neitlier won 
the friction on the dead faces, if it -were (xnittai 
thronghont, and if it acted through the same arcs befo 
and after zero, affect the time directly ; for while tl 
pendulum is falling the friction acta coatiary to grKvit 
and with gravity while it is riaing after the escape h 
taken place. Ab it ia always resisting the motion of tl 
pendulum, it tends to ■^in'l'Il^»ll the arc ; and (m tl 
other hand, the impulse always tends to increase it, i 
that here also there is counteraction to eome extent 
but as the friction on the pallets does not vary in u 
definite proportion to the force of the train, but Bom 
times one way and sometimeB the other, no aaet 
relation of this kind can be established. All we a 
say about the arc ia that it increases under an increaM 
force or a diminished friction, until the remainii 
friction on the palleta and the resistance of the air sto] 
the increase." 

Although the dead-beat is the best escapement fi 
regulators and astronomical clocks, where a weight 
employed as the motive power, it ia not so auitable fi 
smaller clocks with shorter pendulums, or ivhere a mau 
spring is used, as, although its vibrations are iaodmi 
ouB when perfectly made, within centre limits of arc, 
is found impossible to make it so correctly that tbi 
are unaffected by any alterations in the impulse, and 
ia found practically that where auch variations occv 
Uie escapement will gain with any decrease of force, u 
vice versA. The recoil escapement, on the other han 
loses with a decrease of force and gains with an increai 
and as the two conditions counteract one another, it 
the practice to make what are termed half dead-be 


escapements for all but the very highest class of clocks, 
i.e., having a slight recoil, which is just perceptible in 
the movement of the wheel backwards at the resting 
faces of the pallets, instead of making them concentric 
ivith their axis. Tlaa is the best way of making this 
escapement for small clocks in which it is employed, 
which are subject to large variations in the amount of 
impulse ; but I should prefer to rely on a correctly made 
escapement, and a perfect fusee adjustment for equalising 
the impulse, in spring clocks of the better class. 

To draw the escapement (Fig. 66) : — 

From the centre c describe the circle c c', and erect 
the perpendicular p c, this is your line of centres. From 
the same centre draw two radial lines, c a, c a', at 45^ 
on each side of the perpendicular ; the chord of the arc 
of 90 ^^ at the circumference of the wheel will be the 
distance of the pallet centre from the centre c on the 
perpendicular, and the lines I r v^v c' n', will be tan- 
gents to the wheel, at the radial, a c, a' a 

If the pallets are to be circular, mark off 3^ on each 
side of the radials, which will give the width of the 
pallets ; and from p draw the Imes n c p, p / Z', 2^ 
below the line, Z r p, and above the line p c' w', which 
will give the amount of the rest and lift. The angle of 
rest is not shown in the diagram, to avoid confusion, but 
the tooth at r is supposed to be resting ^^ from the 
comer of the impulse plane. 

Draw the arcs r, c, r', c', and join r c, r' c', and 
complete the figure. 

If the lockings are to be equi-distant, mark off the 
width of the pallets 6^ to the right of the radials a c, 
a' c, and proceed as directed. 

In very highly-finished regulators with light trains 
of wheels, the weight driving them is sometimes excessive, 
and the pendulum too light to overcome the consequent 
resistance, which is increasing, at the locking faces of 
the pallets, and as the weight used must be sufficiently 
heavy to overcome any possible increasing resistance of 

308 wATtm Aim clock kakikq. [tnuw^xzi 

the truD, the pendnlam sbonld always be made wi 
ref^race to this, and ehoold alwaya be heavy enon^ 
{Sfivmt tlie nnlockiiig renstanoe haTiiig miult distnrta 
effbct om its ribratioui. 


205. Double Thiee-lefnfed Oravity Eaeapemeni' 
This eecapemeoit wu inTemted bj Sir E. Beckett (then H 
K. B. Deni«on) in 1854. It ie the best eBcapemoLt f 
all very \irge cIocIcb where the hands are exposed to i 
action of the wind, itc, as it admits c£ great dririi 
power being applied without sonmbly affecting tl 
eacapement, by inoreaaed reaistaiice to tite nnlockm^ 
woold be the case with tiie Qraham and other dead-be 
escapements ; and, the impulse to the pendulum bcdi 
given by the weight of IJie arms in falling thiou^ 
given distance to which tiiey have been raised, ii, • 

Fig. 66 is & diagram of thi> eecftpement. a b a, a i 
are two wheels in different pUnes, the lifting pins shci 
at X being pivoted into and between them. The palk 
O R F, u F act between the wheels in the same plM» i 
one another ; the lifting pins acting on the lifting fan 
after the line of centres, on the long teeth or l^s of H 
wheels being released ftom the stops n, k, which tt 
placed one on each side of Uie pallets, and act altemabd 
on the wheela The pallets are pivoted one <ni each W 
of the bending point of the suspension spring. 

The action of the escapement is as follows ^-^V 
pendulum v is travelling in tiie direction o{ tJie amn 
when it has moved the arm H sufBcienUy far, tiia tootii 
will escape from the stop M, and the Ufting pin iriD nu 
tlie arm on thiit side out of contact witiL the peuduln 




until it escapes from the lifting face, when the tooth a on 
the opposite side will fall on to the stop N, the arm if 
falling from where it has 
been raised, and giving 
impulse to iiie pendulum 
at p, until it is stopped 
by the next lifting pin ; 
and the pendulum, con- 
tinuing its course, will 
raise the arm N in the 
same manner. A fly is 
sprung on the arbor to 
prevent jar from taking 
place on the teeth coming 
in contact with the ^' 
stops, the large arch 
through which the wheel 
moves, 60**, at every beat 
rendering it effective ; 
so that in a well-made 
and properly-constructed 
escapement of this kind 
the stopping is quite 
dead, and all danger of 
tripping is avoided. 

In great turret 
docks, where a very 
large fly is necessary, 
the part of the arbor 
on to which the fly is 
sprung is left large, or a 
collet is put on the ar- 
bor. Of course, as Sir 
R Beckett points out, 
any mcrease in the length of the fly renders it more 
effective than an increase in the width. For obvious 
reasons, the fly should be made as light as possible, as, 
indeed, should all the other parts of this escapement. 

Fig. 66.— ^Double Thiee-legged Qiayity 



The wheels and pallets ehoold be cut from sheet ill 
and all the locking faces hardened. In some of the i 
fine escapements Uie fa<«a of the blocks are jewelled. 

The arroe of the pallets below the stops may bi 
any length, but they are generally constructed at 
same angle as the upper arms, as shown in the % 
In large escapements, where they mast be of cooaden 
thickness, and cannot be bent to adjnat the beat^ t 
have adjustable pins or beat screws to oonte in oonl 
with the pendulum. 

To set out the escapement : — With an escape wl 
of a ^ven size, draw a circle representing ita drci 
ference ; from the bottom of the escape wheel, on the I 
of centres, produced, divide this circle into three, ti 
these three points will be the positions of the act 
&ces of the three teeth of either wheel when at r 
and the acting foces of the stops. Draw radii from 
centre of the wheel to these points, and tangents to 
two upper radii meeting above, the wheel will indicate 
theoretical point of suspension and the bending point 
the spring. The angles z A o, o b x will then each = 9 
and the line of centres will ^= the diameter of the wh 
Bloxam, who invented a diflereut form of gravity esci 
ment, made the top portions of the arms cranked i 
pivoted them at this centre. This, however, has b 
found to be quite unnecessary, and they are now mad< 
shown in the figure, safe locking being secured by plac 
the stops higher on the arms than their true tbeoret 
places, which increases the angle, and is, in fact, the tt 
as giving a slight draw on them, as is done in the U 
escapement by undercutting. 

The teeth of the lower wheel are usually set at ec 
distances between those of the upper, although this is 
absolutely necessary, and the lifting pins are set on r 
to the acting faces of the teeth of one of the wheehi 
as to act across the line of centres, at a distance fi 
the centre not exceeding a twelfth of the radius of 
wheel. The lifting pins must be hardened and polisl 


and the lifting faces of the pallets must be either hard- 
ened and polished or jewelled. 

From the comparatively great angle at which the 
arms are to the line of centres, the distance through 
which they have to be lifted to give sufficient impulse is 
less in this escapement than in one with a larger number 
of teeth acting in the same plane, as the pallets would then 
hang more nearly upright. This is a great advantage, as 
the contact is shorter. The unlocking is also easier for 
the same reason, and* from the greater diameter of the 
wheel in proportion to the other parts of the escape- 
ment, the pressure on the stops being considerably less. 

Tlie fly is also more effective than where it has less 
angular motion, which, with the above, renders the 
locking lighter, and, from the few teeth in the escape 
wheel, another pinion is necessary, which reduces the 
effect on the escapement and pendulum of irregularities 
in the train, and the superfluous power before-mentioned, 
necessary for overcoming such defects, carrying the hands 
in all weathers, &c, 

206. Four-legged Gravity Escapement — ^This escape- 
ment is also ihe invention of Sir E. Beckett. It is the 
same in principle as the foregoing, from which it differs 
in the following particulars : — 

It has only one escape wheel, with four teeth or legs, 
and two sets of four lifting pins, one on each side. The 
wheel acts between the pallets, one set of lifting pins 
acting on the one, and the other on the other ; the stops 
point inwards towards the wheel. The wheel turns 45° 
at each beat, and its arbor likewise carries a fly. To 
set out this escapement, draw the circumference of the 
escape wheel ; mark off on this circle 67^^ on each side 
of the line of centres, and draw radii to these points, 
which will indicate the positions of the lockings ; tangents 
to these radii, meeting above the wheel, will give the 
theoretical point of suspension, and, as in the three- 
legged escapement, safety must be ensured by shifting 
the stops a little up and down respectively, l^e lifting 



piss, which act <m the lower pallet face, are planted oi 
radii to the acting faces of the teetli, Uie opposite se 
being midway between them. This secores lifting at tli 
line <^ centrea They must be planted at not loore thai 
a fifteenth of the radios of the wheel from the centre 
The pallets do not bank on the lifting pins as in tb 
three-legged escapement, bot on two banking pins, whid 
should be planted as near as possible to the centre o 
percussion, to avoid vibration, in such a position as b 
allow of a little run taking place before coming ii 
contact with the lifting faces to ensure pnnnpt lifting 
The impulse anna of the pallets may be made parallel b 
the upper arms, and mtli^ longer than shorter. All thi 
parts, wheel, pallets, and stops, must be made as lighi 
as possible, and the contact faces hardened or jewelled. 

I have not given a figure of this escapement, thi 
principle of which may be understood from the fonner 
to which it is inferior for turret clocks, and, aa i' 
requires an extra wheel and pinion, and takes up a Uttli 
more room than the Graham, it has not superseded thai 
escapement for regulator. It may, however, be easilj 
made, and will give good results if the above points an 
carefully attended ta 

207. Electrical Clocks.— In 1840, Mr. Alezandei 
Bain, of Edinburgh, invented a clock to go without weighti 
or springs as the motive power, by the application o 
elec^ioity to the pendulum ; he reversed ^e order o. 
things in the old clocks, aud made the pendulum drivi 
what wheels there were in the train and the hands. 

The pendulum bob was composed (d a hollo* 
cylinder, ooutatning a large coil of insulated wire, thi 
ends of which ran up the rod, terminating in two springs 
upon which the pendulum was suspended, and attached t^ 
which were wires whii^ communicated with a smal 
galvanic battery. 

Two groups of four or five permanent magnets bavinj 
opposite poles were fastened together, each magnet closi 
to, but separated &om, the next, and fixed to the case, thi 


like poles of each group being next to and slightly 
separated from those of the other. 

Over these magnets, but without touching them, the 
bob passed in the oscillations of the pendulum, and the 
galvanic circuit was completed by a sliding bar, which 
rested on the battery terminals ; this bar, being pushed 
backwards and forwards by the pendulum in its oscil- 
lations, made and broke contact with the different poles, 
thus reversing the current at each beat, and rendering 
the pendulum automatic in its action. 

The power was communicated to the wheel work of 
this dock by a propelment in the form of a sort of recoil 
escapement reversed, with a ratchet wheel and click, as 
in all clocks where the motive power is transmitted 
through the pendulum there can be no escapement, 
properly so called. 

Bain's system was much thought of at the time, and 
he carried on a large clock factory at Edinburgh for 
some years, afterwards removing to London ; but his 
clocks were failures scientifically and commercially. Being 
directly controlled by the current from the battery, any 
irregularity in that caused a corresponding irregularity 
in the rate of the clock, the friction of the sliding bar 
also proving a source of failure, and they are now things 
of the past But his system was modified and improved 
by others, and clocks kept going, controlled, and set by 
electricity came into common use. 

About the time of Bain's invention, the late Sir 
Charles Wheatstone invented another kind of electric 
dock, where a motor or driving clock actuates a series 
of dials. 

Mr. Shepherd also spent a great deal of time in 
improving electric clocks, his system being approved and 
adopted by the late Astronomer Royal, as it gave a kind 
of gravity impulse to the pendulum. However, all electric 
clocks that are entirely dependent on the constant action 
of the current will fail sometimes if provision is not made 
to enable the dock to go for a time independent of it. 



208. Jonea'B ByBtem of CoBtrollin^ Clocks. - 
This fact became apparent to Mr. R. L. Jones, vi 

1%. 07.— Jona*^ araUm of CootreDinc ClaAa. 

as manager of ilie mlwa; station at Cheater, had foi 
the importance of true time, and after many ananocesf 
experimenta with electric clocks, he adopted the plan 
controlling a Bain's pendulum applied to an ordim 
clock by cansing the pendulum of an astronomical cL 


to make and break contact with the opposite currents at 
each oscillation, thus transmitting at each beat a positive 
or negative current through the Bain's pendulum or pendu- 
lums, and causing both the normal and controlled 
pendulums to beat in unison. Moreover, the pendulum 
being driven by the clock in the ordinary manner, any 
failure of the electric current did not alSect it for some 
considerable time, provided it did not vary more than 
one beat before the current was restored. 

The following is Mr. F. J. Bitchie's explanation of 
Mr. Jones's system : — 

"Each controlled clock is complete in itself, re- 
quiring periodical winding up, and regulated approxi- 
mately to time, but it is provided with a pendulum, p, 
constructed similar to Mr. Bain's, vibrating over a 
bundle, or rather two bundles, of magnetic bars, SN — NS 
(Fig. 67). The normal clock is furnished wiiii slight 
springs, a and (, one on each side of the pendulum, o, 
which are attached, one to the copper terminal plate of 
battery A, and the other to the zinc terminal of battery 
B ; the other poles, + and — , of each battery passing by 
the line wire or through the earth, to one of the sus- 
pension springs of the controlled pendulum, p, thence 
down the rod and around the baU, R, to the second 
suspension spring, and by the line wire to the pendulum 
rod of the normal clock. 

" When the normal clock pendulum swings and touches 
the one spring, a current of one kind is transmitted from 
the battery, a, along the wire, producing an attraction 
between the one pole of the magnet bar and the wire 
coil of the controlled pendulum, r, the second battery, b, 
being meanwhile inactive until the normal pendulum, in 
its reverse swing, makes contact with the spring 6, and 
a current of the opposite nature is transmitted to the 
controlled pendulum, producing, of course, an opposite 

" Thus each second by day and night an alternative 
current of positive and negative electricity is transmitted 


along the line and through the controlled pendolam 
whidi most effectually secures the coincidence of beat i 
the normal and ooutrolled docks. Should the controlk 
clocks incline to go too slow, theee correcting ouireol 
drag them onwards ; if too &st, tlien the currents retai 
the motion of their pendnlmns, and perfect coincidenc 
of time is secured." 

The success of Jones's plan was at onoe evident, tta 
it was applied to several of the public clocks at liverpoo 
tiiese clocks being controlled hy an aatronomieal elo^ i 
the observatory. 

Some years ago, before any attempt was made t 
controlling, and when the time of tlie public clocks i 
London was very irregular, the council of the Horologia 
Institute sought to indnce the Corporation to establii 
three public clocks, to be controlled on Jones's systen 
Sir George Airy offered to supply tha current froi 
Greenwich Observatory, and Messrs. Walker and Vsrlej 
the eminent electricians, offered every assistance the 
could render. But the failure of a dock belonging t 
Messrs. De la Rue, which was controlled on Jones' 
system, was a sufficient excuse for the Corporation to d 
nothing in the matter ; this clock was, by-the-bye, a ver 
bad one, and has since been removed. 

Inventors and people who are enthusiastic in adoptiq 
an invention are apt to exaggerate its utility and scope 
it was thought that any clock was good enough if it ha 
Jones's controlled pendulum applied to it, and bad clock 
have had a good deal to do witii the partial abaadonmei 
of that system. 

309. fiitoMe'B Electrical Clock Eaoapemeoit. - 
Mr. Bitchie, of Edinburgh, has been the latest an 
most successful improver of electric docks ; by h 
system he is enabled to maintain a number of pendulun 
in motion by currents of electricity transmitted to thei 
by the action of a normal clock, the action being tl 
same as in Jones's clocks, only there are no weights < 
the subordinate ones, Uie vibrations of which a 


maintained hy electricity alonu. As in this case the 
pendnlnm drives the wheels and hands, the accuracy of 
the dock most depend 
to a great extent on the 
oonstmotion of what in 
an ordinary clock would 
. be the escapement, and 
Messrs. Bitchie have in- 
vented a number of inge- 
nioos contrivances for this 
purpose. The escape 
wheel having no tendency 
to move of itself, the pal- 
lets must be so constructed 
that they will hold the 
wheel during a portion of 
the time of the swing of 
the pendulum, and drive 
it forward during the 
remaining portion. 

Fig. 68 shows the 
most approved of Messrs. 
Bitchie's escapements, or 
propelments. It consists 
of a wheel with the usual 
thirty teeth, tlie pallets 
being two gravity arms 
mounted on separate ar- 
bors, and unconnected 
with each other j the long "* 
straight portions of these, 
o, a', act as the crutch. 
The pendulum,?, has just *''»■ **-^^S^iJSf**^ '''"'^ 
completed it« vibration to 

the right and is retoming to the left ; tlie urn of the 
pallet, resting on the tooth at b, is prevented from 
pressing it forwards by the wheel being locked on the 
opposite side by the stop, » ; when the pendulum, 



continuing its vibration, reaches G, it will lift the sttq 
f, oat of the wheel, &nd the arm b, b, already resting ( 
the tooth at b, will press the wheel forwards onlal tl 
next tooth a stopped at >'. 

The projecting plane, a, acts in the same manner i 
the end of the arm at b, preventing the wheel from goii 
baokwarda, and pressing it forwards when it is nlesM 
from the stop, s'. 

For large clocks, where the hands are exposed to tl 
pressure of the wind, a click and ratchet are added to tt 
escape pinion, to prevent the wheel from being 6xw 
backwards when the weight of the gravity arms mi^ 
not be sufficient to do sa 

As clocks with cases sufficiency long for secoiK 
pendulums are sometimes objectionable^ Messrs. Bitchi 
have an arrangement for loading a simple pendolni 
above its point of suspension, by which means a pendi 
lum can be constructed to vibrate within the diamett 
of its own dial to any nte desired, on the principle t 
Maelzel's metronome ; both ends of this pendulum hat 
the bobs coatsining coils of wire passing over msgnel 
as in the longer pendulum, thus securing r^ularity i 
the vibrations, which could not otherwise be depends 

The momentum which these subordinate pendulun 
acquire while they are being impelled by the current i 
sufficient to keep them osdUating for some thirt 
seconds, and consequently, drive the train and hands f( 
that length of time; but if through any accident tb 
motor clock failed to transmit the current, or the cloc 
itself stopped, the subordinate clocks would stop alsa 

Mr, Bitchie's system, however, is in operation i 
many places, and has given entire satisfaction for year 
and he informs me that he has not found it neoeesat 
to make any alteration or improvement in what is herei 

Messrs. Barraud and Lnnd have introduced a systei 
of setting or synchronising clocks by means of cli] 

Chap.ZXiy.] HOUSE CLOCKS. 31 2> 

(actuated by an electric current) catching the minute 
hand at fixed interrals, which is largely in use in London. 
The principle is similar to that applied by Bain, the in- 
ventor of electric clocks, for a similar purpose. It has 
been carried out with great energy by Mr. J. A. Lund, 
and is now worked by a company. 

A good deal of practical difficulty has been expei^- 
enced by the various persons who have tried to set or 
synchronise clocks in this way, from the want of unifor- 
mity in the clocks to be set and other causes, so that, 
whatever may be the advantages of such a system when 
employed in large establishments like the Post Office, 
it is questionable whether in merchant's offices and 
such places a good clock, that can be procured at a 
moderate price, and that will keep time without set- 
ting for weeks and months together, would not be less 
expensive and a good deal more trustworthy than common 
dials set by any automatic system. 



210. Various Kinds of Clocks. — Fifty years ago 
many of the so-called watchmakers in the provincial 
towns were in reality clockmakers, as the names on so 
many of the old long-cased clocks testify. The country 
clockmaker usually had an engine, and could cut his own 
wheels and pinions, but as these were cut with a straight 
slit only, they had to be shaped up afterwards by hand 
with a file. This system was so very primitive that it 
could not last, but it had the merit of teaching the 
country watchmakers' apprentices to file and turn, and 
some of our best watchmakers have been drawn fipm this 



These clocks, having invariablj seconda' pendDlvD 
and long canes, irere not sufficiently cheap or portable £ 
the working populatiooB of large towns, and were mpt 
seded hj the Black Forest, or Dutch, clocks, which 
their turn have given plaoe to the macfaine-made doc: 
of the American factories. Pnusewortb;' eSbrts are no 
being made to reduce the price of English hoose docl 
by adopting machinery largely in their manufactni 
and it is difficult to explain why this has not been da 
befora It ia to be hoped these attempts will lead to ti 
revival of an industry which has been allowed 
dwindle into insigniticance. 

The French have long held the market for omament 
clocks, not because of the clocks being good in themselvt 
but on account of the decorative style of the cases. The 
clocks are generally uniform in size and caliber, an 
although there is veiy little automatic machinery ei 
ployed in their manufacture, are produced very cheapl 

The imitations of the old English bracket clocks ai 
quarter clocks are failures, owing to their small moi 
ments and want of power, btit principally to their lackii 
the adjustment of the fusee, for whatever controver 
there may be respecting the merits of the going ban 
for pocket watches, there is none as to its fitness f 
clocks where all the power obtainable is necessary, and, 
a clock were made to go eight or nine days only with 
going barrel, it would lose seveml minutes in the latt 
days of the week. For this reason the French clocks ■ 
mode to go tweuty^one days, so that if they are wooi 
once a week, a few turns only of the ^>ring are unwonn 
and the pull is kept approicimately equal. But in doc 
requiring greater driving power this long spring cann 
be used, and the unequal pull of the spring can only 
got over by the use of the fusee, which ia one of t 
reasons why we keep the makin g of offioe dials, Ac., 

The Viennese and German weight clocks, introdao 
since the institution of international exhibition^ a 


tolerably portable, and should keep good time, but they 
are not in much demand, as few of them are made to strike 
the hours. If a factory or factories could be established 
here with sufficient capital and machinery, there is no 
doubt we could make clocks cheap enough to compete 
with all comers. I believe that at present only the 
movements of the American clocks are imported, the 
cases and fittings being made more cheaply here; and 
these movements are poor things. However, until the 
factories are set up, provincial watchmakers would do 
well to revive clockmaking, though not exactly on the 
old lines ; and, since the art has to some extent died out, 
I will describe some of the details. 

211. Eight-day Clocks. — The brass work of the eight- 
day striking clock (whether weight or spring) is cast, 
and should be well hammered before anything else is 
done to it, the plates and wheels especially. The wheels, 
and sometimes the plates, are made from hard rolled 
sheet brass, the former being stamped out. The holes 
in the centre of the wheels having been drilled and 
broached to the required size, and the wheels turned true 
on the edge on a screw arbor, they should be filed flat 
on the sides, and of an equal thickness, when they are 
ready for having the teeth cut in them ; but it is better 
(if the proper facilities, such as an engine, exist for 
doing this work) to cross the arms out first, smooth 
the sides, and fix them on their arbors before cutting the 
teeth. The plates, after being well hammered, should be 
pinned together and made square, then separated and 
filed flat, and smoothed by scraping them. 

The pinions of these clocks are now always made 
from pinion wire. Sometimes the centre pinion is cut 
in an engine, but if it is made from the wire, the leaves 
are left on that part of it which is to form the large 
front pivot, and softened and forged until it looks solid ; 
this enables the arbor to be left large enough for a 
shoulder to the pivot. A deep sink is cut on each side 
of the pinion heads, and the leaves are broken ofl" the 




portions that are to form the arbors. This is done 
squeezing two leases together at a time in & strong t 
which eaves the labonr of filing theni off. The aib 
are then centred, and turned and smoothed with a smo 
file. The pinions also requir« smoothing, and often 
bottoms of the spaces need to be sunk a little more wit 
file of the proper shape ; they are then ready for harden 
and tempering. They should be left as hard as il 
possible to turn them, and if in hardening the arbors 
bent, they may be straightened by striking the holi 
side a few blows with the paan of a hammer on a pi 
of steel or iron, care being taken to strike the ar 
opposite to where it is solidly supported. The piuii 
and arbors can now be polished as follows : — 

Cut a square hole in a piece of hard wood that t 
just fit the pinion; grip a board of willow or other b 
wood in the vice, and, with the aid of the piece of h 
wood which forms a couple of handles at right angles 
the board in the vice, move the pinion backward a 
forwards until the leaves have cut out tracks for th( 
selves, and left little ridges between that go to I 
bottom of the spaces. By charging the board w 
emery and oil, and working the pinitm along it, I 
leaves are easily polished, and will not be put out 
shape or rounded. 

The arbors are first polished with oil-stone dust e 
a polisher, and finished by tying two pieces of soft wo 
charged with fine emery, together at one end, a 
gripping the arbor between them while it is rotatii 
This is a very short process; and the pinions can 
faced, if that is thought desirable, either with a squi 
polisher or the ordinary facing tooL The wheels i 
riveted on to brass collets, which are attached to 1 
arbors with soft solder, except the centre wheel, whi 
IS riveted on to its pinion. 

If the depths be now made in a good depthing U 
the caliber can be drawn by marking them off accordi 
to the space at disposal or the size of the frames ; t 


wheels of the striking train are usually placed in nearly 
a straight line, one over the other. The fly should be 
kept quite clear of the pallets, and the centre and escape 
wheels in a line with one another, if the clock is to have 
a seconds hand. Before separating the plates, drill the 
pivot and pillar holes, and open the latter to nearly the 
required size, when the plates can be separated and the 
pillars put in. If the plates are to be lacquered, the 
inside of the pillar plate should be smoothed, and the 
pillars riveted in at once ; but if the plates are to have 
a high poUsh, the pillars must be put in temporarily, and 
should have either a washer under the rivet or a very 
long rivet left The plates are polished by pinning them 
on to a block, and rubbing them smooth with pumice 
stone and plenty of water, and afterwards with Water-of- 
Ayr stone, finishing off with a buff and rotten stone and oil 

English clocks have always the rack spur and lifting 
piece made of steel or iron, but for no adequate reason 
that I can see. In French clocks these pieces are made 
of brass, which is more easily worked, and therefore 
cheaper; and in any rational system they would be 
stamped out. 

212. The Striking Train. — Planting them is almost 
arbitrary. The rack must be planted close to the pivot 
hole of the third wheel in the striking train, the lifting 
piece being in such a position that the end which 
projects through the plate will catch a pin in the 
rim of the fourth or fly wheel The rack hook should 
be planted so that it keeps one or two teeth of the rack 
to the left, or in front, of the third wheel pivot hole, 
and at a right angle to the I'ack. The rack should 
have thirteen or fourteen teeth cut in it and must 
be the segment of a circle, but the first tooth should be a 
little longer than the others, in order that when the 
hook is lifted over that tooth it may clear all the others 
when the clock " warns " (i.e., the noise made by the 
rack arm falling against the snail a few minutes before 
the clock strikes). 




If the clock is required to strike " one " at the hall 
hour, the first tooth must be lower than the second oni 
and the pin in the minute vheel so placed that it wi 
lift the hook over only one tooth for the half-hour. Th 
teetli of the rack should be cut in an engine, but if the 
are marked vith a double-edged spacing punch they t» 
be cut by hand sufficiently accurately ; iu fact, that i 
the method of cutting the teeth on the racks of tb 
finest repeating watches. The snail is usually carried o 
the hour wheel, and the steps on it are marked by the pi 
in the rack arm, the rack spur being shifted from tooth t 
tooth and held tightly in each, while the Bnail is move 
round, and the pin in the rack arm marks it. The ub: 
formity of the teeth of the rack being unimportant b 
long as they coincide with the steps on the snail, tb 
length of the steps is determined by bringing the pii 
In t^e minute wheel against the back of the arm of Ui 
lifting piece, and marking the snail with dividers froi 
any convenient point. The arm of the lifting piece i 
made of hard hammered brass, and should be suffioientJ; 
weak to allow the pin in the minute wheel that lifts i 
from the front side to pass under it if the hands ar 
turned backwards ; the arm can either be twisted, o 
have an ear left on it to enable this to be done. 

The rack arm is also of hard brags ; it should be thin 
and have the pin in it made of such a shape that it wii 
ride over the snail in the event of the clock missing t 
Btrike " twelve ; " the &ce of the sniul should b 
bevelled off for tJiis purpose, as if it does not ride over i 
(Ae clock will stop. 

The rack lifting piece and rack spur have long brae 
collets riveted into them, te which studs are fitted tba 
are screwed into the plates, and are kept in their place 
by small pins put into the studs above the collets. I 
English clocks the hour wheel is not fitted to the stem c 
the minute wheel, but to a bridge that is screwed to th 
pUte over the minute wheel, and as the hour band i 
generally fixed by a small screw to the boss of the hou 


wheel, the wheel should be keyed on to this boss to en- 
able the hour hand to be shifted. 

In clocks with small, or even twelve-inch, dials, I see 
no reason why this bridge should be used ; in French 
clocks it is always dispensed with. If the hour wheel 
were fitted on to the pipe of the minute wheel it would 
improve the appearance of the dial by keeping the 
centres of the hands smaller, and a pipe to the hour 
hand is a more convenient method of enabling it to be 
set with the minute hand than shifting the minute wheel 
in putting the clock together. 

When the clock is required to repeat the hours the 
snail is placed on a separate stud with a star wheel, so 
that, instead of moving gradually, it is shifted at once, 
lust before the hour is struck, in the same manner as in 
a repeating watch, the third wheel of the striking part 
has a large pivot in front, which projects through the 
plate, on which the gathering pallet is squared. This 
gathering pallet in English clocks, in addition to its 
office of moving one tooth of the rack for each stroke of 
the hammer, also stops the train when the rack spur has 
been drawn over the last tooth of the rack, a long tail 
of this pallet then coming against a pin projecting from 
the rack. In French clocks, this stopping is effected by 
the addition of two extra pieces and an extra spring ; 
one of these pieces drops down and catches a pin in the 
rim of the third wheel. Sir E. Beckett says this is a 
better plan than the English one ; it is certainly no 
more effective, however, and is much more complicated 
and troublesome. The train is stopped more easily at 
the radius of the wheel than near its centre, but I have 
never heard of a pivot having been broken off by the 
contact of the gathering pallet with the pin in the rack. 
The old-fashioned clocks had the bell at the top, and the 
hammer stalk parallel to the side of the frame ; and, in 
this case, the hammer spring (usually made of iron) was 
planted inside the pillar plate. The spring had its 
upper end bent or forged into an angle, and a projection 


of the hammer Btalk resting upon the flat head of tl 
spring served as a buffer for the hammer, and prevent 
the JB.rring which would take pla«e if the hammer wta 
allowed to fall against the pUlar. In all clocks ttu 
strike on a bell, spring, or gong, and in bracket cloc) 
where the bell is placed at the ^ck for convemence, tl 
hammer arbor is brought through the back plate, an 
pivoted into a high cock. In this case the hammer 
osoall; placed horizontally, and a spring is screwed o 
to the cock, which comes in contact with the hammt 
stalk, and keeps the hammer free from the bell after 
has struck it. The pin wheel has eight pins set in ii 
rim for lifting. The hammer tail on which these pii 
act must be left as long as possible ; it should I 
shortened until when it drops ofi* one pin it will n< 
quite reach the next, but be a little distance from it Ti 
third wheel must turn once for every pin, or eight timt 
for one of the pin wheel ; and the foiuth wheel must mat 
some equal number of turns for one of the third wheeL 
The numbers of wheels and pinions vary considerabj 
in clocks, but can be easily calculated. (See pp. 57-60 
The train in common use is generally, for the strikin 
train: — 

MainwheeJ. PinwheeO. Thlri. FonrUh. rij. 
WheBl . . 84 64 66 60 — 

PinioQ . . — 8 S 7 7 

Going train, with seconds pendulum : — 

HbId. CsDtn. Tbiid. 

Wheel . . 96 
Pinion . . — 

64 60 30 
8 8 8 

For a clock with half seconds pendulum, a suitabl 

train is : — 

WbMls. Plnioiu. 
84 — 
64 8 
66 8 
60 7 


Wheel!. linlonM, 
96 — 
84 8 
80 8 
30 7 

Chap. X?IV.] HOUSE CLOCKS. 327 

with sixteen turns on the fusee or barrel to allow the 
clock to go and strike for eight days. 

In finally adjusting the striking part, care must be 
taken that the hammer tail when it falls from one pii^ al- 
lows the pin wheel a little run before the next pin comes 
in contact with it, as, if it does not, the clock is likely to 
stop striking when the oil becomes a little thick. In 
putting the clock together, the third wheel pinion must 
be shifted in the pin wheel until the tail of the gathering 
pallet is quite close to the pin in the rack that stops i^ 
otherwise the pin-wheel will have too much run after the 
hammer has dropped. The pin in the fly wheel should also 
be close to the projecting end of the lifting piece, and not 
at the back of it, for the same reason. 

The dial wheels may have any convenient number of 
teeth, that will allow of the hour wheel turning once to 
twelve turns of the minute wheel. The escapements have 
been already so fully treated of that it is unnecessary to 
say anything about thejn here; the half-dead-beat is 
found best for clocks with short pendulums, and the 
recoil is usually applied to those with seconds pendulums. 
It is not usual to make a maintaining power for these 
clocks, but it is necessary if accurate timekeeping is 
aimed at, as, when a seconds pendulum is swinging with 
a recoil escapement, the pallets drive the wheel backwards 
if the power is taken off, and for every second the clock 
should go forward while being wound it will go back one 
second. However, when a clock is thought good enough 
to have a maintaining power it is thought good enough 
also to have a Graham dead-beat escapement, and is re- 
moved from the category of house clocks. 

213. Quarter Clocks are different from those already 
described only in having an additional train of wheels to 
strike the quarters. A good many French clocks, and 
formerly some English ones, were made to strike the 
quarters and hours with the same train. They struck 
usually on two springs, and were called ** Ting Tang " 
quarter clocks; but tbis kind of clock is never made here 




now. The quarter barrel and great wheel innst be large 
than those of the hour part, but the principle of tbi 
striking vork is the same, with rack hook, liftinj 
piece, and quarter snail. The quarter rack ia made b 
discharge the hours. The minute wheel has four pins L 
it, lifting the quarter lifting piece in the usual way ; ani 
when the rack falls into the lowest notch of the snail a 
the fourth quarter, a pin in it strikes the end of the rad 
hook of the hour part, releasing it from the rack, ani 
allowing the rack arm to fall into the hour snaiL Tb< 
quarter rack in falling sets free a lever that is doing thi 
duty of the lifting piece to the hour part ; this lever i 
lifted up by a spring, and the end of it catches a pin ii 
the end of the fly wheel, and prevents the bout' trail 
from moving until the quarters are struck, when tb 
quarter rack, acting on the tail of this lever, draws i 
down again, and releases the pin in the fly wheel of thi 
hour part, thus permitting the hours to be struck a fei 
seconds after the quarters. This quarter train is alwaj: 
placed to the right of the going part, and the uBua 
numbers employed are : — 


Fint wheel. Beonnd. Third. PonrUL Tlj. 

If the quarters are struck on an octave of bells, tin 
pins for lifting the hammetB are set on a barrel in Un 
following order : — 


Chap. XXIY.] 



Thikd Quarter. ^ 






Fourth Quarter. 

If struck on four gongs or bell springs, the Cambridge 
chimes are used. The notes of these gongs must not 
follow one another, as in the bells throughout, but the 
lowest, or fourth, must be a musical fourth below the 
third, and the gong on which the hour is struck one or 
two notes lower. 

The following are the Cambridge chimes, written an 
octave lower to keep them in the lines : — 

First Quarter. 


Second Quarter. @ 

Third Quarter. ^• — |« ~J^ --j— j—] "J * ^ 

!.Z^ feif.^^U -^^ ^ rl, i,i^ii^:z E fl^ 


214. Afitrononiical Clocke or Begulatora buve m 
striking part to them, in order to avoid all complication 
and ae the performance of these clocks depends mainl; on 
their escapements and pendulums, both of which have beei 
treated of under tliose heads, little need be said of theii 
construction. In setting out the caliber, the importanl 
thing to observe is the ptosition of the hands on the dial 
If the dial is a twelve-inch one, as is usual, the escapt 
wheel hole should be marked in a, straight line above tb« 
centre hole, at a distance of 2 '5 inches ; and as the houi 
hand, to avoid friction, is carried on the arbor of a whee 
that coincides with the great wheel, and is geared iuU 
it (or rather, into a smaller wheel with an equal numbei 
of teeth that is screwed on to the back of the greal 
wheel), these wheels must be small enough to allow ol 
the great wheel being planted considerably below th( 
centre wheel, and to the right of it, as then the line wil 
&I1 on the inside of the barrel, and the friction on the 
barrel pivots, as in a left-handed fusee watch movemeni 
(see p. 68), will be the difference between the weight 
and the resistance, instead of the anm of the two. But, 
as the weight must not come down in the middle of the 
case and in front of the pendulum, a barrel with a worn 
cut on it should be placed in brackets, screwed to ami 
projecting from the left-hand side of the frame, at e 
sufficient distance to bring the weight down in the left 
hand comer of the case, and as near the front as possible 
The line should be fixed in front of the barrel, and nol 
nextthegreatwheeI(whichistheusualmetbod),as then the 
weight in descending gets more out of the way of the pen 
dulum as it unwinds the line. The well-known influenced 
swinging pendulums on pendulous weights in proximitj 
to them renders this point of keeping the weight at t 
sufficient distance from the pendulum when it reaches 
the level of the bob of considerable importance to tiu 
good performance of the clock. 

For a regulator of no great pretensioDB, the following 
is a suitable train, with a great wheel of about three 


inches in diameter, the others diminishing propor- 
tionately : — 

Oreat wheel. Centre. Third. Escape 

Wheel . . 144 96 90 30 

Pinion . . — 12 12 12 

They are, however, usually made with higher numbers, 
and, of course, larger wheels, otherwise the teeth would 
not be strong enough. In the best regulators, the 
following is the usual train : — 

Oreat wheel. 




Wheel . 





Pinion . 

• •■^— 




The pallets should be jewelled, but if the pivot holes 
are well made and of the best brass, I see no reason for 
jewelling them, as the objection to long holes in 
watches does not obtain here, and they will last a long 
time without wear. Much of the accuracy of the going 
of the clock depends on the way in which it is fixed. 
The case should have a very solid back screwed to the 
wall some distance from the floor, and the pendulum 
should be hung on a strong cock screwed to the back of 
the case, and in such a position that the bending point 
of the spring will be exactly opposite the pivot of the 
])allet arbor. The crutch should have beat screws, for 
adjusting the beats, and the clock be placed where Uiere 
are no air currents. 

215. Turret Clocks. — ^Until a comparatively recent 
date, public clocks in England did not keep good time, 
the dc^-beat escapement of Graham, that gives such good 
results in astronomical clocks, not being adapted for 
clocks that are subject to external disturbance or have 
a varying motive force. The pin wheel and recoil 
escapements were most commonly used, and various 
forms of remontoirs were from time to time applied here 
to large clocks, and in France, where public time was 
considered of more importance than it was in this 
country, the latter were in common use. The remontoir 



is an arrangement not in the escapement, but in die 
upper part of the going-train, by vbich a weak Bpring ia 
wound lip, or a nmall weight is lifted, that gives impulse 
to the escape wheel at short intervals, by which it was 
sought to counteract the irregularities in the impulse 
that were caused by a coarse train, etc., and the varying 
pressure of the wind, eta, upon the hands. But UieM 
remontoirs were complicated and delicate, and have now 
been entirely superseded by the double- three-legged 
gravity escapement. 

The revolution effected in the clock trade since tbc 
design and construction of the great Weatminster cIoc6 
(1851 — 9) by Sir Edmund Beckett, and the improve- 
ments made by him in the escapement, strikiug parts, 
eta, created a revival on a lai^ scale of the factor; 
system of turret clockmaking in this country, a detailed 
description of which is beyond the scope of the present 
work. To anyone interested in the manufacture of lai^ 
clocks, however, I would recommend a study of ^ 
Edmund's excellent work on the subject 

The numbers of trains, sizes of wheels, fall of weights, 
etc., being governed by the available space and the 
height and formation of the clock tower, the details ol 
the clock should be considered in relation to this ac- 
commodation by the maker, who should so design hit 
clock as to utilise to the utmost the space at his disposal 



ADDENDA of pinion leaves, 50, 51 
Adjusting rod, 83, 93, 106 
Adjustment of chronometers, 32, 

227» 228, 241 

the balance, 32, 227, 228, 211 

chain, 93 

fusee, 16, 63, 65, 66, 71 

mainspring, 66, 71, 78 

maintaining power, 65, 66 

Ahaz, King, 8 

Airej, Sir G. B., 16, 53, 238, 256, 257, 

Alfred the Great, King, 9 
American clocks, 15 
Anchor escapement, 292 — 295 
Angular measurement of lift, 139 
Anne, Queen, 20 
Annual Greenwich trials, 18, 26, 27, 

30. 257, 258 
Applying balance spring, 100, 221, 

t25,226, 227 
Apprenticeship system, 35, 36 
Arbor, Barrel, 73, 88 

— centre wheel, 84 

Arcs of vibration, 18, 25, 209, 211, 

212, 283, 299 
Arnold, John, 24, 155 
Arnold. J. B., 260 
Arnold s chronometer, 24 

— compensation balance, 22 

— <7lindrical spring, 25 

— escapement, 24, 25, 155, 180, 181 
Astronomer Boyal, 23, 29, 313 
Astronomical clocks, 15, 330, 331 
Attaching the dial to a watch, 105 
Attachment of the mainspring, 70, 

72, 74, 75 
Austrian clocks, 15 
Auxiliary comp nsation, 27,242— 257 
Axis of suspension (pendidum), 281 

BAILEY, Francis, P.B.A.S., 288 
Bain, Alexander, 312, 313 
Balance-spring, The, 205—227 

curb-pins, 99, 100, 103, 211 

Cylindrical heUcal, 25, 206, 222 

Balance - spring, the, Blocks for 

making, 222, 223 

Invention of, 25, 208 

Length of, 223, 227 

To apply, 225, 226, 227 

To form the curves, 226 

To harden and temper, 


To polish and blue, 224, 

2 25 

Elasticity of, 18, 206, 210, 213, 

214, 228, 232 

Electro-gnt (Dent's), 212 

Glass, 212, 213, 214 

Gold, 214 

Hand and tool made, 207 

Invention of, 18, 206 

Isochronism of, 18, 25, 206, 208 


Lengths of, 209 

Long and short arcs of, 211, 


Lutz*s, 207 

Materials for, 213, 214 

Number of turns of, 209 

Old-fashioned Swiss stud, 102 

Ordinary volute, 210 

Spring-box and winder 

for, 218 

To apply, 100 

To nardeu and temper, 

218, 219 

To make, 217 

To polish and blue, 219, 

Overcoil, or "Breguet," 102, 


Curve, 221, 222 

Index bad for, 222 

Length of, 210, 211 

Pinning in, 221 

Winder for, 220 

Palladium (Paillard s), 214, 229 

Permanent loss of elasticity of, 

215, 225 

stud and index, 99, 100 

To de-magnetise, 215—217 


Bulow. Edmrd, 1 

lurel vid taaat, *l. *t 

I h[»l!!'70, 72,81,88 


(ecket, air E., 30, S7 


laTDS, Cuibiii ot. 124 

Brtbood, U. 3S, 12B, 

Bmrd oi loogltnde, 31, S3, » 
Bolt ma jirfnt, lOB 
Brunwsir, Sr Fr»derlck, 37 
Bn» sdgs tot diiUa, M 

pivote, 18S, 168 


— — pivoted ud Bprinff det«nta, 1 
Pocket. 188—178 

bu^DS (hon0-Bhofl)t I', 

detent, ITS, 173, 178 

diameter of baJtmce. IW 

— eBCBpe-wheel. 169 

impnlse roOor and ptll 

— — — lockioir-BtoDA 175 
pr^oiiDg the different f«i 

— — — iLnlookiiig--n)llor and p«Ui 

— imlocklug-sprmpr, 175 

Present fonn of, 157 

Proportions of. 158. 180 

— Flniidiinl, SO— 
to adjnit n 

CuBOnjiiiilon. 81, 82, 96, lOS 
Caron, Petar, asa 
Oantra otgatritf, £78 
osdllation, 378. 379 

— puiton, SS, Sj'. 58, 80, 8*. 88, 11 

Cbain, Piuea, 41. 44. 88, 93, 108 

— — E^ly fartaa of, 155 

Earnihaw's. 28, 155 

eecspe-irheol, 183. 183 

~1^T^ ™ ""* 
iQTendon of , 34^ 35, 155 

— Eiglit ind twit4a; nuiina, 58, S 

— moremont making, 37 

— B^Htem of manvfaotnre, 161 

— with eoiog barrel!, 71 
Cironlar error (of peudolnm), 1 

Cit; ud Onllds' Inrtitnto, 18 
Clark, Latimer, 8. 315 
Clement, William, 293 

Kikmakgn- Compuir, 18, 33, 11 



Clocks, eight-day dials, 16, 321—327 

— Electrical, 3^-319 

— French, 14, 320, 323 

— German, 15, 320 

— House, 319 

~ Imitation English, 15, 320 

— Quarter, 15, 327- 331 

— Kegulators, 15, 330 

— Turret, 16, 331, 332 

— at Exeter, 11 
Peterborough, 11 

St. Margaret's, Westminster, 


St. Paul's Cathedral, 10, 11 

St. Paul's, Coyent Garden, 


Westminster, 16, 332 

Club-tooth lever escapement, 188 
Cole, J. P., 186, 187 
Comi>ensated watches, 32,213 
Compensation balance, 22, 227—258 

adjustment, 32, 227, 228 

Arnold's, 22. 25 

Auxiliary, 27, 227—242 

Dent's, a?, 247, 248 

Eamshaw's, 22, 26, 28 

Eiffe's, 29, 244, 247 

■■ Failure or ordinary or " mid- 
dle temperature error," 27, 30, 

Hardy's. 29, 242, 243, 244 

Hartnups, 30, 248—250 

formula, 232 

Invention of, 22 

Kullberg»s, 30, 253-255 

Le Boy's, 22, 25, 252 

Loseby's, 29, 30, 251, 252 

Mercer's, 30, 255, 256 

Molineux's, 29, 244, 247 

Poole's, 30, 250, 251 

to inferior watches, 32 

Theory of, 227-242 

— bar, (Airey's), 256, 257 

Gridiron pendulum, 14, 21, 288, 


— — Mercurial pendulum, 14, 285 — 

— pendulums, 285—292 

Batio ox expansion of metals, 

Wood and zinc, or wood and 

lead, 14, 289, 290 

Zinc and steel, 14, 290—292 

Construction of wheel teeth, 48 

Crank lever escapement, 38, 181 

Crisp, W. B., 212 

Crown wheel escai>ement (cMVerge) 

Crutch (pendulum), 281 

Ctesibius, 9 

Cumming, Alexander, F.B.S., 289 

Curb pins, 21, 99, 100, 102, 211, 222 

Cycloidal theory, 13, 275 

Cylinder escapement (s««Horizontal 

Cylindrical balance sprmg, 25, 26, 


DEAD beat, (Graham) escape- 
ment, 21, 297, ^06 

Charles Frodsham on, 299, 300 


escaping arc, 299 

pallets, 297—299 

, To draw, 307 

De Beaufre Brothers, 110 

Definition of time, 1 

De la Bue's clock, 316 

Deuison, E. B.,16 

Deut, Frederick, 29, 212, 213, 242 

— E. &Co., 212 

Dent's balance, 29, 247, 248 

Depthing, 94, 96 

Detached (chronometer) escape* 

ment, 24, 25, 155-176 
Detent (chronometer), 155, 150 — 161 

— (fusee), 44, 65, 94 

Detents, Long and short, 159, 160 
De Tick's clock, 10, U, 124 
Dial feet holes, 105 

— wheels, 60—63, 96, 97 
Dials, "Eight-day," 16 
Diameter of wheels and pinions, 45, 

Diamantine, 204 
Dillmar, 10 
Dipleido8COi>e, 7, 8 
Dog screw, 110 
Double-bottom cases, 108 

— roUer lever escapement, 176, 177, 

— three-legged grravit^ escapement, 

Drivers and followers, 44, 45, 51 
Duplex escapement, 129, 137 — 154 

Action of, 139, 149 

bankingps, 141 

depths, 140—144, 151 

Drop in, 150—154 

escape wheel, 137, 146, 150, 151 

impulse teeth, 140, 146, 153, 154 

pallet. 140, 150, 152, 153, 154 

in full-plate watches, 145 

intersection of wheel and im- 
pulse pallet, 140, 141—150 

Invention of, 129, 137 

Lift in, 139, 140, 146, 147 

Principle of, 137 

— — Buby roller, 139, 147—160, 152, 

Notch in, 147—150, 153 

— — Bunning of, 141, 145 




Duplex Motpomsnt, Betting of, US 
Tonuke. UO. lU 


Ihitertn, Jwo Baptirte. 137 
Dnt( on lutohsa, UmUd Statae', 


'pABLIEnsUih Blocks (imilstioii), 

Ediptio, The,G, < 
Ednudl.. kW. 10 
1,- dials. If 

■», S7, lei 

r, Hooke'i law ol, 13, 2S 
w Bprini, IB 

sllHkl, Sli-SIB 

B»iB-«, Sia, 31S 

— Ewkpemaut tor, 3te-StB 



En^uh clook tnde, 11, 319—333 

— mtohtnda,S4 

Btl»cloidal teeth, 4S 
Bquldlfltuit lookbinf' tn J«ver es 

capemeot. IDO— ]SS 
EBCBpemenCs (olock), 302-316 

- Daaii tent (GrahRm), 297-Sll 

- De Viotij^lO, 11. la 

EitcMs'B. 317 

— Iwatoh), 133-3C0 

Chronometar, 34, S5, UK-ITO 

DnplBi, 129. 187— IS* 

Horiionbil, 138—187 

Ijsver, 21, SB, 178— a» 

— Teige, 10-13, 133-188 

■pACIO, NicolM, SO, 110, 111, 11^ 

P*!torie>, Wttch, 31 
FflrroHOD, Junes, 7 
FinidiiDB, Utmnometar, M-SS 

— WntoS, 8S-10a 

— wliselB uid pliiiiHH, 33 
Fitting watch liends, IM 
FlUng movement in case, 106 
FlBola ol wfaeel't«eth and pinion- 

luTea, 50, SL 

Fog'e patent mnion, 50 
Foraerr ol BngliHb Hall niai 

311. 3ll """ ' ""f*" 
Freeuig burol, SO 
French clocks, 11. 40 
Friction, 11, IS, «, 188 

— on esiape-wheel teeth, 188, 18 

— OD rneee pilots, SB 
Frodehom, Charles, 37, SIS, 

298.239. 181,308 

on dead-beat eKwemeDt 


Full-jOate watchaa, 39, 10, 11. 61 
Fnaae, Adjustment. 16, 3^ (B, 71 

— ■nd barrel {size of). 11, 83 

— cap and hook, Bl. 67, B7, 93 

— Nnmberor turna on, m,ee 

— pieoe in top plate. S7, 83 

— atop, 87. ^ 63 

— To tree great wheel OD, 37 

— To [jant, 88 

QALELEOfl disooTor of 

Galileo, VincaDio.'l2, 374, 37B 
Oanees, 8ir J. Whitworth'B, 37 

— Want or (tandatd, 37 
Oearing of wheels, 14—53 
-White's, 48 

Generating oirdea, 15. 48, 17 
Oeneta atop work, 75—77 

German olocks, 14, 19 
Qilding, 120-133 
Gimbela (or ehroi 


Ooing-liarael. 71, 73, 75—78 
Ootluo vault, 48 

Graham's meronrial peudoJun, 

IraTity, Centre of, 278 
— esoapemaita, SI, 906—313 

Action ^ 308, 309 

Double thiofrlegged, 308- 

ronr-loggad, 811^318 

To set out, 310. 811 

Sreat Bheelofwatiih, 144 65-80, 



Greenwich obserratory, 22. 23, 29, 

— Meridian of, a— 6 

— Standard dock at, 290 

— time, 8 

— trials, 18, 26, 27, 80, 257, 258 
Gridiron pendolom, 14, 21, 288, 289 
Grignion, Thomas, 275 
Grossmann, Moritz, 190 
Guildhall Library and Mnienm, 19 
G, Yalae of, 277 

H ALLEY, Dr., 21 
Hall marks in cases, 84 
Hands, To fit, 61, 104 
Hard brass, 42 

Hardy's balance, 29, 241, 243, 244 
Harrison, John, 20—23 
Harrison's compensation, 21, 28, 

— chronometers, 20, 21, 23 

— escapement, 22 

— maintaining power, 23, 63 

— tomb, 23 
Harris, Bichard, 274 

Hartnup, Professor, 27, 28, 232, 236 
Hartnnp's balance, 248, 250 

— formula, 232, 239 

— tables, 237, 238 
Hautfenille, L'Abb^, 179, 206 
Hele, Peter, 17 

Helical, or Cylindrical spring, 25w 
26, 206, 222-227 

— teeth, 47 

Hengham, Sir Balph de, 10 
Herschen, Sir J., 37 
Hewitt, Thomas, 36, 263 
High numbered moyements, 55 
Hooke, Dr. Eobert, 18, 19, 25, 137, 

Hooke's law, 18, 25, 206 

— recoil escapement, 292—295 
—Hooking in nuiin spring, 80, 81, 89 
Horizontal, or <nrlinder, escape- 
ment, 126-137 

Action of, 127 

adaptability to going-barrd 

watches, 184 

Brass escape wheel. 20, 129, 181 

Diameter of cylinder, 131 

Drop on cylinder, 131, 133 

Impulse curres of teeth. 131, 


— — Invention of, 20, 126 
Lift, 181 

Lips of cylinder, 131 

Overbanking in, 127, 129 

Pivoting in cylinder, 135 

— — Flagging cylmder. 186 
Proportions of, 183 


Horizontal escapement. Baby <qrlin- 
ders, 20, 129 

Thickness of cylinder, 131 

Tool for measuring heights, 


To harden and temper cylin- 
der, 136 

To make cylinder, 136 

To plan, 130 

To run m cylinder, 134 

Horologe, horologium, &c., 9 

Horological Institute, 19, 29, 35„ 
217. 232, 316 

—Journal, 22, 232 

Hours a watch will go, 59, 60 

Hour wheel, 62, 96, 105 

House clocks, 319—331 

— of Commons, 24 

Htmtsman's steel, 172 

Huyghens, Christian, 12, 13, 18, 179, 
206, 274, 275, 293 

Hypocydoid, 45 

TlOnSCH, Moritz, 209 

■*• Improyed chronometer balances* 

— going barrel, 76, 77 
Index, Balance spring, 99 
Involute teeth, 47 
Interchangeable movements, 42 
International ezhibitians (1851 aad 

1862), 207, 238 

Commissioners of, 207^ 206 

Isochronism of the balance sraingt 

18, 206, 203 
; oonditionB of, 18, 206, 


pendulum, 12, 18, 275 

Italian computation of time, 8 

TAMES n., King, 19 
^ Jewel holes, 113, 116 

— screws, 98, 116 
Jewelling, chronometer, 84 

— hole making, 118—116 

— Invention of, 20, 110 

— Objections to, 110 

— Stones used for. 111 

— Watch, 110 . 
Jodiu, Mons., 1J» 

Jones', B. L., dock, 814^ 116 
Julius CsBsar, 8 
JOrgensen, A., 71 

XTATEB'S (Captain) p6ndulum,279 
-^ Kendall. Mr., 23 
Keyless watcn. Napoleon's, 259 
Madame de Pompadour's, 256 



GtaVoB^ m-vt 

— —ToMMitn—Vi 

— — EninwK'i, S7«-i73 

' nHkiDKbar, SI,S6S-SM 

XnUbDrg'! a>t.Tim talan«, SSS 
•- Impiond udinuT UUuMe, SH 

L'XBSB BnaMsQillg, 17B, 3M 
Ijmb. Bllmiln, 280 

It^ato rigtouL # 

Lau«L iluiM of pinJoB. M, 47 

— To and Dnmben ol, Wj^W 
Lift-huided mora 

I^^ (pm.rt 

Id Bar, Jnllen, 19, 2K 2M31 

nerra, 1S9, wTuS, iOS, 162 

Lnor (HapeuoA >^ 3^ ITS-auS 

leUon of, 17«-17S 

buklna plBi, 178, SOS 

dnmlki uOIeti, 190— IH 

etab-tootb, 189 

cnok lerer, 38, 181 

depth!, aw, aos 

deUohed (Madge'*), 2i, 88, 180 

donbls rolla, ITS— 178,181, ISt, 



- -• lockiuKS, 17B, IBS, 102, 19S 

- - Bi>t6bia laiHT, 301 

Ovrrhankiiig in, 178. 188, 187 

F^let tUKiealn, 18^ 185 

- tepellaut(Cole'el,lB7 
-resiUent (Cola'e),18S 

- roller, ISO, ZOO, SOI, 305 

- senii-e^cular psUeU, IOS 

Able roller, 38. 182 
V) make, lOB^^to 

^ — horiioutAl ee 

™tl», I' 



Litlierlud, Fatar, __. 
Livopool obvervatorj, 

— wUch tnde.3S 
Lo«-aimib«ied pmicma, 5i 
Losety'a babuice, 28. 2si, 352 
XoufitadCL BauiIo(,ai, 33.U 

— Keridi^ot, 8 

— BewiwdB for ducoTerj of, 21, ' 

Lnti*! laluoe apdngi, DOT 

UA.CBISS tooiL u, m, ta 
Hniiuprfiw, The, 80-80 

— idjiutnient, 70, 71. 78, 100 

— fttUchment. 70. 71, 7^ 71, 7S 
—hook. 70, 73, 89, SO, 91, 02 

— Hookms iB, 80, 81, 8B, to, M, 

— InTention at, 17 

-^ Len^tli of, for gDlsg-burelB, 

— MbJeui^, 78 

— KnmbBT ol ooili In, 72, JT 
^Pivoted bfmoe for, 73 

— tepared, 70. 71, 79 

Mukaljiie, Dr. HeiiL a 

VtMaj, Thomu, S^ ISl 
Hau Hlu time, 6 
MediDi, Prinoa Lewcddo da, IS 
Hercer'* litUiiaoe, 30. 81. 3SS, iSS 
UaroDTial glldiDg, 131. 133 
— pendBlnm, H ffiS— 288 

Hinute wheel, 62, 98 
Molioeoi'a balauiAe, 2a, lU, 
Momaot of inartia, 220 

— wheeli, m-ia, te, 97 
HotamentB, Kuiae chnn 


— Watoh, 87— M 

fnU-plata, », ID, 41, ei 

Hi^-nimbared, U 



Morements, Watob, Lancashire, 87 

Left-baixded, 68 

Machine-made, 37 

Plates for, 42 

Three-quarter plate, 40, 61 

Madge, Thomas, 23, 24, 68. 120, 180 

his chronometer, 23, 24 

— leyer escapement, 24, 38, 180 

-VTAPOLEON'S keyless watch, 259 
■^ Nelthropp, Bey. H. L., 275 

his treatise, 276 

Nickel movements, 121 
Number of coils in mainspring, 72, 77 
Numhen of watch trains, 53--63 
NnrendMrg watches, 17 

riBSEBYATOBY of Greenwich, 22, 

^ 23,291,816 

Liverpool, 27, 8D 

Ogive, 48 

Oil, Watch, 112 

Old clock at Exeter, 11 

— Peterborough, 11 

St. Paul's Cathedral, 11 

Oovent Ghu-den, 274, 275 

Westminster, U 

^enburgh, Mr., 206 

Old-fashioned spiinff stud, 

Orolo^ario, Bartholomo, 10 


Oxtexjt Lord, 9 

Osoilmion, Centre of, 278, 279 

Oscillations, Number of, in p«nda- 
Ixuns, 877, 978 

Orerbaiiking in horizontal escape- 
ment. 129 

Overooil (Brsgnet) spring, 102, 206, 

PAILLABD, Monsieur, 214 
Palladium springs, 214, 229 
Pallets, Chronometer, 161- 163, 169 

— Duplex, 140, 150, 152—154 

— Lever, 190—195 

— Verge, 124, 125 
Pendulum, The, 274—292 

application to clocks, 12, 274, 


arc of vibration, 283, 299 

Barometric error of, 283 

Centre of gravity of, 278 

oscillation of, 278, 279 

Circular error of, 13, 276 

— — crutch, 281 

— — Cycloidal theory of, 18, 275 

Pendulum, The, DisooveiTol, 12, 275 
formula for finding time of vi- 
bration, 276 
•^ — grravitation pendalnra, 279 

Isochronism of, 12, 18, 175, 136 

necessity of compensation, 14 

principle of compensation, 14 

(000 Compensation pendulums) 

regulation, 284 

sbape of bob, 283 

sidereal seconds, length of, 279 

Simple, 278, 279 

Suspension of, 279—281 

To nnd lengths of, 277 

Philippe, Mons., M 
Phillips, Mons., 206 
Pillar phites, 48, 44 
Pinion, Cannon, 61, 62, -06 

— Centre, 57, 58, 60, 84, 86, 107, 108 

— Fog's patent, 50 

— nuudng, 42, 43 

— wire, 42—51 

— Wheels and, 44, 52 
Pinions, Addenda of, 48—51 

— depths, 94 

— Facing, 84, 322 

— Lantern, 4^ 

— of high and low numbers, 49 

— Polishing, 84, 322 

— shape of leaves, 45—48 

— To find numbers of, 57 — 63 
Pin-wheel escapement (Lepaute's), 

Pitch circles, 45, 5Q 
Pivots, shape of balance staff, 165, 

166, 211 
Pivoted brace in going barrel, 72, 


— detent, 155 
Pivoting In wheels, 94 
Planetariums, 9 
Planting wheels, 81 
Plugging cylinder, 136 
Pocket chronometers, 169 — 176 
Polishing pinions, 84 

— wheels, 83 

Pompadour, Madame de, 258 
Poole's balance, 30, 250, 251 
Pope Sylvester IT., 10, 124 
Potance, 39, 40, 126 

Power exerted by wheels on pinions, 

Prescot^— 40 
Prest, Thomas, his keyless work, 

Proportion of adjusted watches, 32, 


QUABE, Daniel, 19 
Quarter clocks, 15. 327-^1 




-oaoUuUdii. ffi 

xpied of wheel! In tnin, E7, S 

Beooil, or wcbor, MCkpaDant, 201- 

Bandpitan, IS, 190, SSI 
BMd, TbomM, la, n,M 
BApAiitiiiff work, '" 
Rppe1l3ndflTar I 
BHUiest-lerer I 
R«Blattnoe of Ixmma, oa 
BI«K, EdwK^ MJ., MO, BJl 
Hitoble'i slectriaa clooki, 318 

— loftded pflndDliUlU, SiB 

— prowlTnent, 317 
Bo^w£, H., 71 

BookinB-lHr kejleai work, SI, BS3— 

Boiler, Duptai, 13S, 117— UD, 1S3, 

— Lerer, ]7S, 199-301, 305 

BniuLliiff of dnplox 
Bab7 OTlInflsn, 30 

C AltD^LAfiSEa, 9 

c? annniei. CUndlm. 71, 73, ! 

B&>a«e, Qeorge, 31, 183 

— PollshTng, BS 
~ Taps, M. 117 
_ TemperiDK, 9S 
Set-liaiid piece, 104, 107 
ShikBpeie, 17 
Bhepheid, Vllliun, S49 
ehlp'H GhroaonietaTm, 30-31 
Siderwl olock at Oreenwicb. E9D 

— Bflconda' peDdnlnm, 379 
Simple 'peidolnm, 13, 373, 378 

SmWton. John. F.B.S., SM 
Smuling, 1 IS— 120 
— l«rrel (u-bor, 73, 88, 89 
Bocietfor Arta,a9 

Speoinl loan oolleetdon, SenMngUm, 

SteadTpinniig, IM, 197 
Steel, Hontamui'H, 171 
St. Kutueri rWertmiiwtai) OoO, 

St. Piol'B Cathednl eloek, 10. 11 

Suipension of pondulmim, 379— 3B1 
BwlH levn watelm. 190 
STnobronlBed elockfl, 318, 319 
BfUoOarn., Pope, 10, 131 

rriABLE of erpuudon of meta], 383 
■*- TiIbl»Ioller lerer eaapenwt, 

8S 18B 
Table! tor oonctmotlan at UttA- 
Tabl« I "™"" 

rHHtnop'B), 237, 338 
Tables, Clarke'i tranitt, S. 3 
Teetli, Bpiojoloidal, «, 46, 47, 1 

— FbiakB df, go 

— Hjporarelofdal, «, 4S, 47 

— Bbainot, U— IS 

— To find nomberH oL 57— 83 
ThermometO' kerbe (Barrlsciii' 
Tbinning wbeels, 83. 86 
Time, Andent Eoman, 1 

— Chinese, 3 

— Deflnitian i^, 1 
-Equation of, a 

— Hean sDlar, S 

— Sidenal,3 

— Bolat or difl, S 

Timiiw wljostmenta of cb 
meters. SI, S3 

— •onwi.Sl, S3 



Tool for taking heights in chrono- 
meters, 1G4 
Tools. Machine, 41, 42, 4.3 
Tompion, Thomas, 19, 126 
Tracing a meridian, 6 
Trains, 53 

— Duplex, 53, 141 

— Past and slow, 54 — 57 

— Horizontal, 54 

— Lever, 54 

— Marine chronometer, 56 

— Nnmhers of, 43, 54 -CO 

— Pocket chronometer, 53 

— Bmdstance of, 53 

— To calculate train, 57 

— Verge watch, 54 
Transit instrument, 2, 3 
"Treatise on Modem Watchmak- 
ing," 124 

— Eev. H. L. Nelthropp's, 275 
Turns, Number of, for fusee or 

barrel, 59, 60 
Turret clocks, 331 
** Two-day " marine chronometer, 

57, 161 
Two-pin lever escapement, 183 
Tyrer, Thomas, 137 

UNITED States' duty on watches, 
Units of weights and measurement, 

Uprighting in mandrel, 88 

VALUE of G, 277 
Varley, Walker and, 316 
Verge escapement, 10, 11, 12, 124— 


Action of, 124 

angle of pallets, 124, 125 

Vibration, Arcs of, 18, 25, 124, 209— 

213,283, 299 
Vibrations, Number of, 53, 54 
Viviani, 12 

Volute, ordinary balance spring, 210 
Vulliainy, B. St. Just, 280, 281, 300, 
. 302 

TyALKEE and Vurley, 316 
'» Want of standard gauges, 37 

of time, 17 

Watch and chronometer movement 
muking, 37^41 

— examining, 103—108 

— finishing, B5— 1()2 
- Riding, 120—122 

— je velliug, IM— 118 

— trade, Euirliah, 3 1, 35 

in Birmingham, .'14 

Clerkeiiwell. :Ji. 

Coventry, ;j4 

Liverx>ool, 34 

Watches and chronometers, 17, 

Water cIock><, 8 
Wesrm lister clock, 10, 332 

— oldclo* k, 11 

W'-eatstoue, S r Charles", o7, 313 
Wheel, Arms, 13 

— Hour, 62, !)(3 

— Minute, «»-*, 96 

— teeth, Sliai)0 f, 4j — 49 

Construction of, 48 

F auks of, 50 

Generatiui? circles for, 45, 40, 


To find nnmhers of, 57—63 

Wheels, To olish, 83 

— To th u, 8:J, 86, 

— Trains of, 53 

— and pinions, 41 — 52, 94 

Bevelled, 51, 52 

depths, S4 

Diameters of, 49—50 

I »ivi'ling off, 51 

Gearin^f of, 47 — 49 

Lead of, 48 49 

Motion, 60—63 

Pitch circl.s of, 45, 48—51 

White'. urst, John, 281 
White's gearing, 48 
Whitwor h. Sir Joseph, 37 

— gauges, 37 
Willis, Pr«'fes.-or, 49 
W'ndiug squares. 107 
Wriglit, T. D., 210 
Wycherley, John, 36, 41, 43 







For some time past there has been a widespread demand on tlie 
part of technical students for text-books. The object ot tbii series 
is to meet this demand by furnishing books which describt ike 
application of science to industry. 

These manuals are not intended to teach pure science, nor 
yet the methods of a workshop ; they are to form the link which 
will combine these two essentials for the modern artisan — thought 
and practice ; they are designed to make workmen thinkers, not 
merely human tools. 

To attain this, it is most important to translate the language 
of the Professor into the language of the workshop, thus giving 
to the British workman an acquaintance with the scientific prin- 
ciples which underlie his daily work. 

No special knowledge of mathematics or of science is 
necessary to the student of this series, but it is expected that he 
will have been observant of the processes carried on in his work- 
shop, which will be here scientifically explained. The subjects 
will be treated analytically rather than synthetically ; that is to 
say, the machine, as the workman knows it, will be taken as a 
whole and analysed, and special care will be taken to avoid the 
method so conmion in scientific books, according to which a 
number of abstract principles are first developed, while their 
practical application is deferred to the end of the book, which, 
probably, the practical man never reaches. 

As text-books for the large and increasing number of candi- 
dates at the Technological Examinations of the City and Guilds of 
I^ndon Institute, with which many of the authors are connected, 
these Manuals of Technology will be especially valuable. An 
author in each case has been selected who was able to comprise 
a well-grounded scientific knowledge with a practical familiarity 
with the minute details of the trade treated of in his book. Con- 
sequently, while the latest and most approved processes of manu- 
facture will be found described, the exact scientific reasons for the 
superiority of these modern methods over the older ones will be 

it; •>! 

1 ' 

Hi I 


: ir 


given in full, as well as such indications as science would suggesi 
for improving the present processes. 

List of Manuals of Technology. 


Watoh and Clock Making. By D. Glasgow, Vice 

President, British Horological Institute. Price 4s. 6d. 

Design in Textile Fabrics. By T. R. As hkn hurst 

of the Technical School, Bradford. With Coloured Plates 
Price 4 s. 6d. 

Spinning Woollen and Worsted. By \\ . s 
Bright McLaren, M.A. 4s. 6d. 

Steel and Iron. By W. H. Greenwood, F.C.S. 
Assoc. M.l.C.E. 5s. 

Practical Mechanics. By J. Perry, M.E., Professor o 

Mechanical Engineering and Applied Mathematics, City anr 
Guilds of I^ndon Technical College, Finsbury. Price 3s. 6d. 

Cutting Tools Worked by Hand and Machine 

By R. H. Smith, Professor of Mechanical Engineering, Sii 
Josiah Mason's College, Birmingham. Price 3s. 6d. 

Fluid Motors. By Professor Perry, M.E. 

Electric Lighting and Transmission of Power. 

By W. E. Ayrton, F.R.S., Professor of Electrical Engineering 
and Applied Physics, City and Guilds of London Technica 
College, Finsbury. 

Chemistry. By Dr. Armstrong, F.R.S., Secretary of the 
Chemical Society, Professor of Chemistry, City and Guildi 
of London Technical College, Finsbury. 

Flax Spinning. By David S. Thomson, Manager, 

Mountain Mill, Belfast. 

The Dyeing of Textile Fabrics. By Professor J. J 

Hummel, of the Yorkshire College, Leeds. 

*»* Other Volumes will be added. 

The numerous Illustrations in these books arc pictures o 
the actual machines as they exist in the best factories, and no 
conventional representations conveying but little intelligence to th< 
practical man. 

The aim throughout has been to prepare books that shall appea 
at once to the workman. Their preparation has been entrusted t( 
writers who know what the workman's difficulties are, what infer 
mation he needs to help him in his trade, and this is presented ii 
such a form that the reader may be attracted by a desire to lean 
the Why and Wherefore, instead of being repelled by the suppose( 
difficulties of science. 

Cassell & Company, Limited : London, Paris & New York 

Selections from Cassell <fc Coinpaufs 


Algebra, The Elements of. By Prof. Wallace, is. 
Animal Painting in Water Colours. With i8 

Coloured Plates by Frkdkrick Taylp:r, and Instructions. 5s. 

Arithmetic Examples, The Marlborough. 3s. 
Arithmetic Rules, The Marlborough, is. 6d. 
Arithmetic, The Modern School. By George 

Ricks, B.Sc Lond. In Seven Books. Standards I. to IV., 
pap)er covers, 2d. each ; cloth, 3d. each. Books for Standards 
V. to VII., paper covers, 3d. each ; cloth, 4d. each. Answers, 
6d. ; or the Complete Set in One Vol., with Answers, 2s. 

Arithmetical Test Cards, The Modern School. 

By George Ricks, B.Sc. Lond. For all the Standards. 
Seven Sets of 40 Cards, 9d. each set ; or linen lined, is. each set. 

China Fainting. By Florence Lewis, of the Laml)eih 
School of Art. With 16 Coloured Plates, and a selection of 
Wood Engravings. With full Instructions. 5s. 

Commentary for Schools. Being the separate Books of 
the New Testament Commentary for English Readers 
(Edited by the Lord Bishop of Gloucester and Bristol), in 
Small Volumes, suitable for Educational purposes : — 

An Introduction to the New Testament, 2s. 6d. ; St. Mat- 
thew, 3s. 6d ; St. Mark, 3s. ; St. Lukb, 3s. 6d. ; St. John, 
3s. 6d. ; The Acts of the Apostles, 38. 6d. ; Romans, as. 6d. ; 
Corinthians I. and II., 3s. ; Galatians, Efhbsians, and 
Philippians, 3s, ; CoLOSSiANS, Thessalonians, and Timothy, 
3s.; Titus, Philemon, Hebrews, and James, 3s. j Peter, 
JuDE, and John, 3s. ; The Revelation, 3s. 

Decisive Events in History. By Thomas Archer. 

With Sixteen Original Illustrations. Seventh Edition, Extra 
fcap. 4to, cloth, 3s. 6d. ; or, cloth gilt, gilt edges, 5s. 

*' These turning-points of history are well told and admirably illus- 
irated. ' ' — Educationa I Times. 

Decorative Design, The Principles of. By 

Christopher Dresser, Ph.D. With Two Coloured Plates 
and numerous Illustrations. Clieap Edition, Cloth, 5s. 


How to Draw Elementary FormB, Models, &c. 120 

Studies in Freehand. By Charles Kyak. Price 6d. 
How to Draw Floral and Ornamental Forms. 80 Studies. 6d. 
HowtoDrawIiandsoapes, Trees, Ships, &o. 6eStudies. 6<1. 
How to Draw Animals, Birds, and Dogs. 64 Studies. 6d. 
N.B. — The above Books contain Blank Paper for copying each example. 

Cassell & Company, Limited : London, Paris & New York, 

Catstll i: Company's Educational Warii — coHliniud. 

Superior Drawing Books, 

riBtedb Fat-iimili by Llihociaphy, ue sl» publiibEd, buund in clatl 

How to Dtaw S'lBnre*. 96 Giuluaied Snidfes from Life. 
Hov to Draw in Vreeluiid. 96 Gnuluued Studiis of Modeb 
How to Drkw rlor&l and Veietable forms. g6 Siudii 

nliakl Dri 
A. T. Elk 

Bnoyolopeedio Dictionary, The. A New and Origins 

Work of Reference to oil Ihe Words in the English Language 
with a Full Account of their Origin, Meaning. Pronunciation 
and Use. Seven Divisional Volumes now ready, price los. 6d 
each, or Double- Diiisionnl Vols, bound in half-morocfo 

EngliBli Literature, Iiibrary of. Edited by Professoi 

Henky Morlev. With Illustrations taken from Original MSS.. 
fie. Complete in 5 Vols., each complete in itself, £a i8s. 6d, 

EnKlish Literature, Diotionary of. Being a Compre 

fiensive Guide to English Authors and their Works. By W. 
Davenport Adaus. 720 pages, exiia fcap. 4to, cloth, 7s. 61I 

DRlish Literature, A First Sketch of. B) 

Henry Morlev, Professor of English Litetatute at Universit) 
College, London. 912 pages, ci^wn Svo. Ttalk Ediliim. 
Cloth. 7s- 6d. 
Euclid, Cassell'S. Being the Firet Six, with the Eleventli 
and Twelfth Books of Euclid. laand Thousand. Edited b) 
Prof. Wallace, M-A. Nem Editien. Qoth Ump, is. 

Euolid, The First Fotir Books of, for the use ol 

Primary Sohools. Paper, 6d. : cloth, gd. 
Figure Fainting in Water Colours. With 1! 

Coloured Plates by BLANCHE Macabthur and JENNIE 

Moore. 7s. 6d. 
Flower Fainting in Water Colours. First and 

Second Series, each containing so Fac-simile Colourai 

Plates by F. E. Hui.ME, F.I_S., F.S.A. 55, each. 
French -English and EagllBh-Frenoh Diotionary, 

Cassell's. i861h Thousand. Crown Svo, (,150 pages, ftoth, 
3s. 6d. : or in superior binding, with leather back, 45. 6d. 

Cassell & COUFANY, Limited ; London, Pakis ft New York, 

Cassell ds Company* s Educational Works — contimied, 

French, Cassell's Lessons in. Nno and Revised 

Edition. 107th Thousand. Revised by Prof. E. Roubaud, 
B.A. Parts 1. and II., cloth, each, 2s. 6d. ; complete, 4s. 6d. 
Key, is. 6d. 

Gkklbraith and Haughton's Scientific Manuals. 

By the Rev. Professor Galbraith, M.A., and the Rev. Pro- 
fessor Haughton, M.D., D.C.L., F.R.S. Cloth, red edges. 

Arithmetic. Containing nearly 

2,000 Examples. 3s. 6d. 
Plane Trigonometry. 2s. 6d. 
Euclid. Books T., II., III. 2s. 6d. 
„ IV., v., VI. 2s. 6d. 
Mathematical Tables. 3s. 6d. 
Mechanics. 3s. 6d. 
Optics. 2s. 6d. 

l-lydrostatics. 3s. 6d. 

Astronomy. 5s. 

Steam Engine. 3s. 6d. 

Algebra. Part I., cloth, 2s. 6d. 
Complete, 7s. 6d. 

Tides and Tidal Ourrents. New 

Edition, with Tidal Cards, 3s. 

Geometry, Cassell's Course of Practical. By 

Ellis A. Davidson. Consisting of 64 Cards in Packet, em- 
bracing the First and Second Grade Studies. In Packet, 5s. 

German-English and English- German Pronoun - 
oing Dictionary, Cassell's. With an Appendix con- 
taining the most important changes introduced into German 
Orthography in 1880 by the Prussian Minister of Education. 
Crown Svo, 864 pages. 46th Thousand. Cloth, 3s. 6d. 

German Beading, First Lessons in. By Fraulein 

Jagst. a Graduated Coiu-se of Lessons in Verse and Prose, 
with Vocabularies, &c. Illustrated throughout, is. 

History of England, The Little Folks'. Ninth 

Edition. By IsA Craig-Knox. Illustrated. Cloth, is. 6d. 

Latin-English and English-Latin Dictionary, 

Cassell's. With brief Illustrative Quotations, Mythological 
Notices, and Short Biographies. 48th Thousand. 3s. 6d. 

Marlborough Arithmetic Examples. 3s. 
Marlborough Arithmetic Rules, is. 6d. 
Marlborough French Exercises, The . Tenth Edition, 

Enlarged and Revised, By the Rev. G. W. De Lisle, M.A. 
Cloth, 3s. 6d. 

Marlborough French Grammar, The. Seventeenth 

Edition, Enlarged and Revised. Arranged and Compiled by 
the Rev. J. F. Bright, M.A., Master of the Modern School 
in Marlborough College. 2s. 6d. 

Marlborough German Grammar, The. Fifth and 

Enlarged Edition. Compiled and Arranged by Rev. J. F. 
Bright, M.A. Cloth, 3s. 6d. 

Mechanics, The Practical Dictioijary of. By 

Edward H. Knight, Civil and Mechanical Engineer, U.S. 
Containing 2,000 Drawings of Machinery, Instruments, and 
Tools in use by every Profession and Trade, with Comprehen- 
sive and Technical Description of each subject. Yowx Vols., 
cloth, jf 4 4s. 

Cassell & Company, Limited ; London, Paris & New York. 

Caiiiii A Company's Ediitalienal Works — 

New Code of Begulations, Handbook of. 

John F, Moss, Oerk of the Sheffield School Board, is. 

Painting in Ifeutral Tint, A Course of. w 

Twenlv four Plales by R. P. I.kcTch. With full Inslruclii 
to the Pupil 5S. 
Phrase and Fable, A Dictionary of. By the H 

Dr. Bkewer. Giving the Derivaiion, Source, or Origin 
nboul aaooo Common Phrases, Allusions, and Words t 
have a Tale to Tell. Sixlientk EdilioH, Enlarged. Clo 
3s. 6d : half-leather, 4s. 6d. 
Physioal Science, Intermediate Text-Book < 

By F. H. Bowman, D.Sc, &c Illustrated. 3s. 6d. 

Popular Educator, Caasell's. An Encvclop^dia 
liLKMENT.VHV AND Atjvancei) Euucatcon. Complete 
six Vols., cloth, price js. each. 

Soienoe for All. Completu in Five Vols. Edited 
Dr. ROBKHT BkowN, M.A.. F.I..S., &c., assisted hy Emin 
Scienlilic Writers. Each containing about 350 lllustmiio 
Extra crown 4to, cloth, 9s. each. 

Sepia PaintinK, A Course of. With 24 Plates fr 

Designs by R. P. Leetch, 4to, cloth, 55. 
Sliakspare Reading Book, The. Being Seventeen 
Rhakspere's Plays Abridged for the use of Schools and Pu 
Readings. By H. Covrtkope Bowen, M.A. lUuslial 
Crown 8vo. cloth, 3s. 6d. -, or in Three separate Books, ts. ea 

Spelling, A Complete Manual of, on the Princi) 

of Contrast and Comparison. By j. D. Mokell, LL. 
H,M. Inspector of Schools, /Slh Thousand. Cloth, is. 

Teohnioal Educator, Cassell's. An Encvcix>p^] 

OF Technecal Education. Complete in Four Volun 
With Coloured Frontispiece and numerous Illustrations 
each Volume. Extra crown 4(0, 416 pp.. 6s. each : 
Two Volumes, half-calf, 31s. 6d. 
Tree Painting in Water Colours. By w. H. j. Bo 

With 18 Coloured Pliies, and valuable Tnstructions by 
.^riisL Oblong crown 410, clolh, 5s. 

Water-Colour Painting, A Course of. Nnn . 

Enlarged Edition. With Twenty-four Coloured Plates fi 
Designs by R. P. LEtTCH, and full Instructions to the Pu 
Clolh, 55. 

Cassell & Company, Limitbd; London, Paris & NewYo 

I ■ ■ •««■ 


Published by Cassell & Company, Limited. 

English History, The Dictionary of. A Record of 

the Doings of the English Nation at Home and Abroad. 
Royal 8vo, cloth, 21s. 

Italy: from the Fall of Napoleon I. in 1815, to the Death of 
Victor Emmanuel (of Savoy), First King of United Italy, in 
1878. By John Wkbb Pkobyn. Demy 8vo, cloth, 7s. 6d. 

Humphry Sandwith. A Memoir by his Nephew, 
Thomas Humphry Ward. With Portrait. 7s. 6d. 

India: the Land and the People. By sir Jam is 

Caikd, K.C.B., F.k.S. With Map of India, and numerous 
Full-page Illustrations. RevUed and Enlarged Edition. Demy 
8vo, cloth, los. 6d. 

John Bright, Life and Times of. By W. Roher ison. 

608 pages. With Portrait. Extra crown 8vo, 7s. 6d. 

Oliver Cromwell: The Man and his Mission. 

By J. Allanson Picton, M.P. With Steel Portrait. 7s. 6d. 

Constitutional History and Political Develop- 
ment of the United States. By Simon Sterne, of the 
New York Bar. 5s. 

A History of Modern Europe. By C. A. Fyffe, 

M.A., Fellow of University College, Oxford. Vol. I. 12s. 

Fine-Art Library, The. Edited by John Sparkks, 

Principal of the South Kensington Art Schools. Vax\\ con- 
taining about 100 Illustrations : — 

The English School of Painting. 
By M. Ernest Chesneau. 
Translated by L. N. Ethering- 
TON. With an Introduction by 
Prof. RusKiN. 5s. 

The Flemish School of Painting. 
By A. J. Wauthrs. Translated 
by Elise Rossel. 5s. 

Artistic Anatomy. By M. Mat- 
thias Duval. Translated by 
F. E. Fenton. 5s. 

The Dutch School of Painting. 
By M. Henry Havaru. Trans- 
lated by G. Powell, (/m pre- 

A Prospectus post free on application. 

Vignettes from Invisible Life. By John Kadcock. 

With numerous Illustrations specially executed. 3s. 6d. 

A Hide to Khiva. By the late Col. BURNABY. New and 
Cheaper Edition, is. 6d.; People's Edition, 6d. 

The Life of the Rt. Hon. W. B. Gladstone. By 

George Harnett Smith. Popular Edition, cloth, 3s. 6d. ; 
Jubilee Edition, with Portrait, is. 

Russia. By D. Mackenzie Wallace, M.A. Cheap 
Edition, in One Vol., with 1 wo Maps, los. 6d. 

Cassell & Company, Limited : London, Paris & New York. 

I IWunui />tibliihcd hy Ceusell &• Cemfany, Limited — cenlim 
UuiTersal History, Caasell's Illustrated. Vui,. 

11.. t 111. Wilh Dumetous high-class EngraWngs, 9s. es 

I OleaningH from Popular Authors. Cunuilett! 

Tno \"cj1s. Willi Original llluslraiions. Price 9s. <s\Kh. 

England, Cassgll's History of. With about 2,1 

Tllustralions. Nine Vols., tlolh. gs, ench. Ot in lite 
bindinf. /Q\ los. A Supplemeniary \'olumi,-. bringing 
1 Work down lo dale, 9s. 

' United States, Cansell's History of the. With ( 

llluslralioni and Maps. 1,950 ijiges. Complen: in Tli 
Vols., doth, j£i 71. ; or in library trinding. £1 ros. 
India, Casaell's History of. With al>out 400 Mn 

I'hms, and Illustrations. Extra crawn 4to, Two Vols., tk 
18s, : Two Vols, in One. ijs. : or in library binding, £i. 
The BusBO<TarkiBh War, Casaeirs History < 

Complete in Two Vols, With about 500 llluslrations. 
each. Library binding. Two Vols, in One. 15s. 

British Battles od Land and Sea. By Jamks Gra> 

Willi about 600 Illustrations. Three Vols., tlolh, £\ 7s.; 01 
library binding, £1 10s. Supplementary Volume, fi.ibrac 
Recent Bniish Bailies, cloth, 9s. 
Old and New Loudon. A Narrative of lis History, 
People, and iis Places. Wilh 1.200 llluslrations. (■mpl 
in Six Vols. , 95. each ; or in library binding, ^3. 

Old and New Edinbursli. By Jamks Grant. \V 

numerous Illustrations, specially executed for the Work, fr 
Sketches iiiadp on the spot, and Aulhenlic MS.**, Complete 
Three Volumes, extra crovra 4to,.cloth, 9s. each. 

Our Own Country. An Illustrated Get^raphicil a 
Historical Description of the Chief Places of Interest in Cir 
Britain and Ireland. Complete in Six Volumes, containi 
about r,aoo Illustraiions and Six Steel Plates. 7s. 6d. each. 

Heroes of Britain in Feaoe and War. By 

HouDKR. With 300 Illustrations. Library binding, Two Vi 
in One, 10s. 6d. ; Popular Edition, complete in Two Vo 

NTOTICE. — A Claaaifled CaUlaCna. giving full parlicalan 

Cuaell A OOmpUiy'l PnbllCatiOlU, raneimg in prici /' 

SIXPENCE TO TWENTY^FIVE GUINEAS, orlll 6r unt an rr^' 

'm lo any addrtu. It wiU U found oftlu trtatisl coHVimt 
sc viko inaji *j trlteliHg hnis /or Sftciti XaxdiHg. Sdimilia 

'rtd Eookt so arrt^Mgrd as to skov at a gtaiKt ikt various Wo 
I valuailr Sr/rctioH, tokich tan ic frocund al Ikt firicet nan 
Booksttlirs- and at Ikt Bookslath. 
■ Riqmtl/or CatalogKt ikoa/d ie addnacd lo 


Li^att Hit/. Liiidi,-,